Automatic slag salvaging method and system for electrolytic aluminum tank and computer readable storage medium

Abstract

The invention relates to the field of electrolytic aluminum unmanned intelligent factories, in particular to an automatic slag salvaging method and system for an electrolytic aluminum tank and a computer readable storage medium. The automatic slag salvaging method for the electrolytic aluminum tank comprises the following steps: controlling a laser radar to globally scan the electrolytic tank, and positioning the position information of the electrolytic tank through feature extraction; a circulation track is generated according to the position information, the crown block is controlled to move according to the circulation track, meanwhile, the first scum thickness in the multiple electrolytic cells is obtained through thickness measuring equipment on the crown block, threshold value judgment is conducted on the first scum thickness through a preset thickness threshold value, and a target electrolytic cell is obtained from the multiple electrolytic cells; scanning the target electrolytic cell, segmenting the point cloud of the liquid level area, and extracting a second scum thickness through the point cloud of the liquid level area; and inputting the first scum thickness and the second scum thickness, and obtaining the corrected thickness according to the first confidence coefficient of the thickness measuring equipment and the second confidence coefficient of the laser radar.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of electrolytic aluminum unmanned intelligent factory, and particularly relates to an automatic slag salvaging method and system for an electrolytic aluminum tank and a computer readable storage medium. BACKGROUND

[0002] In electrolytic aluminum production, aluminum oxide dregs (anode dregs) are continuously generated in the electrolytic tank. If not cleaned in time, it will hinder ion migration, destroy the balance of the electric field and thermal field in the electrolytic tank, reduce the electrolysis efficiency, increase energy consumption, and even cause abnormal operation of the electrolytic tank, affect the quality of aluminum products, and even lead to production stoppage.

[0003] With the continuous development of automation technology, some automatic devices for electrolytic aluminum slag salvaging have appeared on the market. These automatic slag salvaging devices are mostly based on the thickness of aluminum oxide dregs to trigger the slag salvaging action. Usually, a laser device is used to measure the thickness of aluminum oxide dregs, and when the measured thickness reaches a pre-marked threshold, the slag salvaging device will automatically enter the tank to perform the groove salvaging operation. However, due to the complex environment in the electrolytic tank, there are factors such as steam and dust interference, and the laser device is prone to inaccuracy when measuring the thickness of aluminum oxide dregs, resulting in a deviation between the measured result and the actual thickness. Moreover, the existing slag salvaging device usually runs according to a pre-set program, and almost every time there will be an over-salvaging phenomenon, which will bring out the original electrolyte, further leading to a great waste of energy. SUMMARY

[0004] To solve the problems of inaccuracy of laser measurement of aluminum oxide dregs thickness, over-salvaging of the existing slag salvaging device leading to energy waste and abnormal operation of the electrolytic tank, the present application provides an automatic slag salvaging method and system for an electrolytic aluminum tank, and a computer readable storage medium corresponding to the method.

[0005] The application provides an automatic slag salvaging method for an electrolytic aluminum tank, which comprises the following steps: controlling a laser radar to globally scan the electrolytic tank and positioning the position information of the electrolytic tank through feature extraction; generating a circulating track according to the position information, controlling a crown block to move according to the circulating track, and respectively acquiring the first dross thickness in a plurality of electrolytic tanks through a thickness measuring device on the crown block, and performing threshold judgment on the first dross thickness through a preset thickness threshold, so as to acquire a target electrolytic tank in the plurality of electrolytic tanks; scanning the target electrolytic tank and segmenting the liquid surface area point cloud, and extracting the second dross thickness through the liquid surface area point cloud; inputting the first dross thickness and the second dross thickness, respectively acquiring the corrected thickness according to the first confidence of the thickness measuring device and the second confidence of the laser radar, selecting one of the first dross thickness, the second dross thickness or the corrected thickness as the working thickness according to the range where the first confidence and the second confidence are located; calculating the slag salvaging motion path according to the position information of the target electrolytic tank and the working thickness, and controlling the crown block and the slag salvaging shovel of the crown block to perform slag salvaging.

[0006] In a further aspect of the application, the first confidence is dynamically evaluated based on a measured standard deviation and a signal-to-noise ratio, and the second confidence is dynamically evaluated based on the quality of the scanned point cloud.

[0007] In an optional aspect of the application, inputting the first dross thickness and the second dross thickness, respectively acquiring the corrected thickness according to the first confidence of the thickness measuring device and the second confidence of the laser radar, and selecting one of the first dross thickness, the second dross thickness or the corrected thickness as the working thickness according to the range where the first confidence and the second confidence are located, comprises: after normalizing the first confidence and the second confidence, judging whether any one of the first confidence and the second confidence exceeds a first threshold; if the first confidence exceeds the first threshold, taking the first dross thickness as the working thickness; if the second confidence exceeds the first threshold, taking the second dross thickness as the working thickness; and if neither the first confidence nor the second confidence exceeds the first threshold, taking the corrected thickness as the working thickness.

[0008] In a further optional aspect of the application, the automatic slag salvaging method for the electrolytic aluminum tank further comprises: taking the n times of measurement values most recently collected by the thickness measuring device, calculating the standard deviation of the n times of measurement values and the signal-to-noise ratio of the n times of measurement values; inputting the standard deviation into a first quantization function to obtain a first confidence contribution value, wherein the first quantization function is monotonically decreasing with the increase of the standard deviation and always positive; inputting the signal-to-noise ratio into a second quantization function to obtain a second confidence contribution value, wherein the second quantization function is a tangent function; and combining the first confidence contribution value and the second confidence contribution value to obtain the first confidence.

[0009] In a further optional aspect of the present application, the automatic drossing method of the aluminum electrolysis cell further comprises: dividing the liquid surface area point cloud into equal-sized grids, counting the number of points in each grid and calculating a second standard deviation; marking each grid with a number of points less than a number threshold as a low-density area, and setting the second confidence level to be lower than a preset confidence value if at least one of the number of low-density areas is greater than a number threshold or the second standard deviation is greater than a standard deviation threshold, the preset confidence value being less than 1 minus the first threshold; and obtaining the second confidence level based on an ideal standard deviation and a standard deviation under a preset scenario if there is no at least one of the number of low-density areas being greater than a number threshold or the second standard deviation being greater than a standard deviation threshold.

[0010] In a further optional aspect of the present application, the scanning the target electrolysis cell and dividing the liquid surface area point cloud and extracting the second dross thickness from the liquid surface area point cloud comprises: scanning the target electrolysis cell to obtain a point cloud set; extracting a local point cloud under a preset height range from the point cloud set, and performing plane fitting on the local point cloud to obtain the liquid surface area point cloud; and taking the liquid surface area point cloud as a reference surface to obtain the second dross thickness.

[0011] In a further optional aspect of the present application, the drossing motion path is calculated according to the position information of the target electrolysis cell and the working thickness, and the drossing is controlled by the crane and the drossing shovel of the crane according to the drossing motion path, which comprises: positioning the initial position of the drossing shovel according to the position information; obtaining a target height according to the working thickness, controlling the drossing shovel to descend from the initial position, and adjusting the speed of the crane by PID according to the current height and the target height during the drossing descent; and wherein the control strength of the drossing shovel is mapped according to the working thickness when the drossing shovel is drossing.

[0012] In a further optional aspect of the present application, the automatic drossing method of the aluminum electrolysis cell further comprises: after the drossing is completed, controlling the gravity sensor to measure the dross amount, and scheduling the crane to dump the dross according to a preset dross box position; controlling the crane to return to the target electrolysis cell again, and controlling the thickness measuring device to measure the first dross thickness again, and controlling the laser radar to scan again to calculate the second dross thickness; performing threshold judgment on both the first dross thickness and the second dross thickness, and if both are less than a thickness threshold, the corresponding target electrolysis cell is completed; and if any of the two is greater than the thickness threshold, the drossing motion path is continued to be generated and the drossing is controlled by the crane and the drossing shovel of the crane.

[0013] In another aspect, the present application also provides an automatic drossing system of an aluminum electrolysis cell, comprising: a crane; a laser radar connected to the crane; a thickness measuring device connected to the crane; and a controller electrically connected to the crane, the laser radar and the thickness measuring device, and configured to execute the automatic drossing method described above.

[0014] Compared with the prior art, the present application has the following beneficial effects:

[0015] This application controls a laser radar to globally scan the electrolytic cells and uses feature extraction to locate their positions. Based on this information, a looping trajectory is generated, and the overhead crane is controlled to move along this trajectory. This looping trajectory design enables the overhead crane to traverse all electrolytic cells in an orderly and efficient manner. The crane uses a thickness gauge on the overhead crane to obtain the first scum thickness. Simultaneously, the target cell is scanned and a point cloud of the liquid surface area is segmented to extract the second scum thickness. Using two different measurement methods and equipment to obtain scum thickness information, the system can assess scum thickness from multiple angles and dimensions, avoiding the errors and limitations of a single measurement method and improving the accuracy and reliability of thickness measurement. After inputting the first and second scum thicknesses, a corrected thickness is derived based on the first confidence level of the thickness gauge and the second confidence level of the laser radar. Based on the range of the first and second confidence levels, either the first, second, or corrected thickness is selected as the working thickness. This confidence-based processing method fully considers the reliability of different measurement equipment and measurement results, effectively reducing the impact of measurement errors on scum removal operations and making thickness estimation more robust.

[0016] By accurately measuring the slag thickness and selecting the appropriate working thickness, the actual condition of the slag in the electrolytic cell can be accurately understood. During slag removal operations, calculating the slag removal path based on the accurate working thickness ensures that the slag shovel operates at the appropriate depth and range, reducing unnecessary material loss and equipment wear, and avoiding problems such as over- or under-removal due to inaccurate slag thickness assessment.

[0017] Other features and advantages of the embodiments of the present invention will be described in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the specific implementation methods or the description of the prior art. Obviously, the drawings described below are some implementation methods of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without any creative work.

[0019] Fig. 1 A flow chart of an automatic slag removal method for an electrolytic aluminum cell provided in an embodiment of the present application;

[0020] Fig. 2 This is a schematic flow chart of step S40 in the automatic slag removal method for an electrolytic aluminum cell provided in an embodiment of the present application;

[0021] Fig. 3 A schematic diagram of the modules of the automatic slag removal system for the electrolytic aluminum cell provided in an embodiment of the present application. DETAILED DESCRIPTION

[0022] The terms "second direction", "first direction", "third direction", "inside", "outside" and the like that appear below to indicate directions or positional relationships, unless otherwise specified, are to be understood as being based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they should not be understood as limiting this application.

[0023] Furthermore, the use of "first" or "second" in describing features is for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features identified. Features identified as "first" or "second" may explicitly or implicitly include at least one of the identified features. The use of the word "plurality" generally implies at least two, such as two or three, unless otherwise specifically defined.

[0024] In this application, unless otherwise specified or limited, terms such as "mounted," "connected," "connect," and "fixed" should be interpreted broadly. For example, they can refer to fixed connections, removable connections, or integration; mechanical connections, electrical connections, direct connections, or indirect connections through an intermediary; and internal connections between two components or interactions between two components. Those skilled in the art will understand the specific meanings of these terms in this application based on the specific circumstances.

[0025] In the description of this specification, if the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" appear, it means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and the features of different embodiments or examples without contradiction.

[0026] Reference Figs. 1-2 In order to solve the problems in the prior art, the present invention provides an automatic slag removal method for an electrolytic aluminum cell, comprising:

[0027] Step S10: Control the laser radar to globally scan the electrolytic cell and locate the position information of the electrolytic cell through feature extraction;

[0028] Step S20, generating a circulating track according to the position information, controlling the crown block to move according to the circulating track, and respectively acquiring the first dross thickness in the plurality of electrolytic cells through the thickness measuring device on the crown block, and performing threshold value judgment on the first dross thickness to obtain the target electrolytic cell in the plurality of electrolytic cells;

[0029] Step S30, segmenting the liquid surface area point cloud in the target electrolytic cell, and extracting the second dross thickness through the liquid surface area point cloud;

[0030] Step S40, inputting the first dross thickness and the second dross thickness and obtaining the corrected thickness according to the confidence, and selecting one of the first dross thickness, the second dross thickness or the corrected thickness as the working thickness according to the range where the confidence is located;

[0031] Step S50, calculating the movement path of the dross according to the position information of the target electrolytic cell and the working thickness, and controlling the crown block and the dross shovel of the crown block to perform dross according to the movement path of the dross.

[0032] First, control the laser radar to globally scan the electrolytic cell, use the laser radar device to scan the entire working environment, obtain the point cloud data of the surrounding environment, and then locate the position information of the electrolytic cell through feature extraction in the point cloud data: process the scanned point cloud data, and identify the features of the electrolytic cell through feature extraction algorithms such as edge detection or shape matching suitable for the shape of the electrolytic cell, so as to determine the specific position information of each electrolytic cell in the working environment, including the center coordinates, the electrolytic cell size, etc.

[0033] According to the position information of the electrolytic cell obtained by positioning, a circulating track is automatically planned with the shortest distance as the optimization target, and the track can make the crown block pass through each electrolytic cell in turn, and control the crown block to move according to the circulating track.

[0034] The skilled in the art can understand that in some embodiments, the fixed track can also be directly generated based on the existing electrolytic cells according to the internal environment of the factory, but considering that in the process of producing electrolytic aluminum, the number of electrolytic cells in the factory may be adjusted according to the production plan, equipment update and other factors, for example, adding new electrolytic cells to increase production, or eliminating old electrolytic cells. With the method of generating corresponding track by itself, when the number of electrolytic cells in the factory changes, the system can automatically adapt to such changes and generate appropriate tracks without the need for large-scale reprogramming or complex manual intervention. In contrast, the method of generating fixed tracks directly based on existing electrolytic cells may need technicians to re-set and debug the tracks on site after the number of electrolytic cells changes, which consumes a lot of time and labor costs, and when the slag skimming system needs to be applied to a new factory environment, the layout of electrolytic cells and site conditions may differ in different factories. The method of generating tracks by itself can quickly and accurately generate slag skimming tracks that adapt to new scenarios according to the position information of electrolytic cells in the new factory environment, making the method highly versatile and scalable.

[0035] The above-mentioned circulating track can refer to a cruising route, and the crown block is timed or continuously patrolled in the factory according to the cruising route.

[0036] During the movement of the crown block according to the circulating track, the thickness measuring device (such as a plurality of ultrasonic thickness gauges or a plurality of laser ranging sensors) installed on the crown block is used to conduct mobile measurement on each passing electrolytic cell to obtain the initial thickness data of the dross at different positions in each electrolytic cell, i.e. the above-mentioned "first dross thickness". It can be understood that the thickness measuring device is usually provided with multiple devices to obtain more data to avoid errors.

[0037] A threshold value of the dross thickness is set, and the first dross thickness of each electrolytic cell is compared with the threshold value. If the maximum value of the first dross thickness of a certain electrolytic cell exceeds the threshold value, the electrolytic cell is determined to be a target electrolytic cell, i.e. an electrolytic cell that needs to be treated by slag skimming.

[0038] Segmenting the liquid surface area point cloud in the target electrolytic cell: for the determined target electrolytic cell, the point cloud data of the liquid surface area of the electrolytic cell is obtained again by using the laser radar or other appropriate sensors, and then the point cloud of the liquid surface area is separated from the overall point cloud data by a point cloud segmentation algorithm; the second dross thickness, i.e. the thickness information of the dross, is calculated based on the segmented liquid surface area point cloud; unlike the first dross thickness, the second dross thickness is calculated based on the segmented point cloud.

[0039] The first scum thickness and the second scum thickness obtained are input into a processing model, which combines the first scum thickness and the second scum thickness, and the first confidence of the thickness measuring device and the second confidence of the laser radar, to calculate a corrected thickness. At the same time, the processing model will select according to the range where the first confidence and the second confidence are located; specifically, one of the first scum thickness, the second scum thickness or the corrected thickness is selected as the working thickness according to the range where the confidence is located.

[0040] According to the position information of the target electrolytic cell, the working thickness is calculated to plan the movement path of the crane and the scum shovel: according to the accurate position information of the target electrolytic cell and the determined working thickness, the movement path of the crane and the scum shovel is planned, which needs to ensure that the scum shovel can accurately reach the position of the scum in the target electrolytic cell, and considering the thickness of the scum, the descent depth and movement trajectory of the scum shovel are reasonably set to ensure that the scum can be effectively scooped up to avoid excessive labor.

[0041] According to the movement path of the scum, the crane and the scum shovel of the crane are controlled to scoop the scum: the planned movement path information of the scum is transmitted to the control system of the crane, the crane is controlled to move according to the planned path, and the action of the scum shovel is accurately controlled to scoop the scum according to the predetermined trajectory and depth, so as to complete the cleaning work of the scum in the target electrolytic cell.

[0042] In summary, the general inventive concept of the present application controls the laser radar to globally scan the electrolytic cell, locates the position information of the electrolytic cell by feature extraction, generates a circulating trajectory according to the position information of the electrolytic cell, and controls the crane to move according to the trajectory. The design of the circulating trajectory enables the crane to orderly and efficiently traverse all electrolytic cells. The first scum thickness is obtained by using the thickness measuring device on the crane, and the second scum thickness is obtained by scanning the target electrolytic cell and segmenting the liquid surface area point cloud. Two different measurement methods and devices are used to obtain the scum thickness information, which can evaluate the scum thickness from multiple angles and dimensions, avoid the errors and limitations that may exist in a single measurement method, and improve the accuracy and reliability of thickness measurement. After inputting the first scum thickness and the second scum thickness, a corrected thickness is obtained according to the first confidence of the thickness measuring device and the second confidence of the laser radar, and one of the first scum thickness, the second scum thickness or the corrected thickness is selected as the working thickness according to the range where the first confidence and the second confidence are located. The processing method based on the confidence fully considers the reliability of different measurement devices and measurement results, effectively reduces the influence of measurement errors on the scumming operation, and makes the thickness estimation more robust.

[0043] By accurately measuring the thickness of the dross and reasonably selecting the working thickness, the actual situation of the dross in the electrolytic cell can be accurately mastered. In the dross salvaging operation, the dross salvaging path is calculated based on the accurate working thickness, so that the dross salvaging shovel can work at a suitable depth and range, reduce unnecessary material loss and equipment wear and tear, and avoid the problems of excessive dross salvaging or insufficient dross salvaging caused by inaccurate judgment of the thickness of the dross.

[0044] Based on the above overall inventive concept, an example is that: an electrolytic aluminum workshop has 10 electrolytic cells (numbered 1-10), which need to be recognized by an automatic dross salvaging system when the thickness of the dross exceeds the threshold value, and the path of the overhead crane for salvaging dross is planned. The known parameters are as follows:

[0045] The laser radar precision is ±2cm, the scanning range is 10m×10m, the ultrasonic sensor used by the thickness measuring device has a precision of ±1mm and a sampling frequency of 10Hz, the number of sensors on each overhead crane is 8, the dross thickness threshold value is 5cm, the maximum speed of the overhead crane is 1m / s, and the acceleration is 0.5m / s 2 ; the dross salvaging shovel adopts a similar jaw type, and the overall size of the double jaws is 1.2m in total length and 0.8m in width.

[0046] The point cloud data of the electrolytic workshop scanned by the laser radar (such as about 100,000 points) is input; adaptive threshold segmentation (threshold value d=0.05m when distance D=5m) is used to divide the point cloud into three categories: ground, electrolytic cell and obstacle.

[0047] The surface plane of the electrolytic cell is fitted by the RANSAC algorithm, and the parameters are: iteration number 1000, distance threshold value 0.03m. The electrolytic cell edge straight line is identified (Hough transform, angle resolution 1°, distance resolution 0.01m), and the positioning result can be output as: electrolytic cell 1 center coordinates: (2.3m, 4.1m, 0m), length 6m, width 3m; electrolytic cell 2 center coordinates: (8.7m, 3.8m, 0m), length 6m, width 3m;…other electrolytic cells are similar to the above.

[0048] The center coordinates of electrolytic cells 1-10 (numbered) are input, and the global path planning is performed: the A* algorithm is used to generate the shortest path, and the cost function is the distance and the turning cost (turning angle weight 0.2). In addition to pursuing the shortest total length of the path, the present application also tries to reduce the turning angle of the overhead crane during movement, because a larger turning angle may cause the overhead crane to move unstably, reduce efficiency and other problems. By reasonably setting the turning cost weight, a balance can be achieved between the path length and the turning difficulty, so as to generate a more reasonable and efficient path.

[0049] In a specific way, the cost function f(n) = g(n) + h(n) evaluates the path: where g(n) is the actual cost (such as distance) from the starting point to the current node; h(n) is the estimated cost from the current node to the target (commonly Manhattan distance or Euclidean distance), add a turning angle weight in g(n), that is, if the current direction and the last step direction are different, add a fixed weight value, avoid frequent turning.

[0050] The center coordinates of several electrolytic cells are input into the cost function for evaluation, and the path is randomly generated to calculate g(n), turning cost, and h(n) of each path, update f(n), select the node path with the smallest f(n), if the node path is the end point, backtrack the path, end the output loop path, specifically, the core is to add a turning weight to g(n) to make the path algorithm try to go straight, and the open list is implemented by priority queue (heap), and the direction change is checked during traversal.

[0051] Example path: electrolytic cell 1→3→5→7→9→2→4→6→8→10. In the process, obstacles (such as maintenance tools) are detected in real time, and if the obstacle distance <1m, the dynamic window method (DWA) is used to adjust the speed. The path points are fitted with a quintic polynomial to ensure continuous curvature; example: path points (2.3, 4.1)→(3.0, 4.5)→(3.5, 5.0) are fitted as: x(t) = 2.3 + 0.7t - 0.5t2 + 0.2t3 y(t) = 4.1 + 0.4t + 0.1t2 - 0.05t3 (t∈[0,1]).

[0052] Subsequently, when the vehicle moves to above electrolytic cell 5 on the same day, the ultrasonic sensor measures the data (100 samples in 10 seconds); wavelet denoising is used, and the calculated mean μ = 5.12 cm and standard deviation σ = 0.08 cm are obtained, if μ > 5 cm and σ < 0.1 cm, it is determined that the target electrolytic cell is determined. (Threshold judgment as described above), the first scum thickness of electrolytic cell 5 is output as 5.12 cm (marked as the target electrolytic cell).

[0053] Control the laser radar to scan electrolytic cell 5 again, and the liquid surface area point cloud (about 5000 points) is used to fit the liquid surface plane using RANSAC, set the iteration number and distance threshold, and extract the top several scum top points through point cloud segmentation, calculate the average distance of the several scum top points to the liquid surface plane. Example: the average point coordinate Z value of the several scum top is 0.6048, the liquid surface height z = 0.6 m, and the second scum thickness is 4.8 cm.

[0054] Then, the first scum thickness and the second scum thickness are inputted, the corrected thickness is calculated according to the first confidence and the second confidence, and the first scum thickness, the second scum thickness or the corrected thickness is selected as the working thickness according to the range of the first confidence and the second confidence, for example, 5 cm. The center coordinates of the electrolytic tank 5 and the corresponding working thickness 5.12 cm are inputted. The scum collecting movement path, i.e., the movement path of the scum collecting shovel and the clamping / release movement of the scum collecting shovel, is determined. The crane movement is controlled according to the scum collecting movement path, and the scum collecting shovel of the crane is controlled to collect the scum.

[0055] Based on the above general inventive concept, the first confidence is dynamically evaluated based on the standard deviation and the signal-to-noise ratio of the measurement, and the second confidence is dynamically evaluated based on the quality of the scanned point cloud.

[0056] It can be understood that the first confidence reflects the stability and noise level of the measurement data. If the standard deviation is large or the signal-to-noise ratio is low, it indicates that the data reliability is poor, and the system can dynamically reduce the dependence on such data of the thickness measuring device to avoid misjudgment of the path due to noise (such as misjudgment of the position of obstacles). On the contrary, the quality of the point cloud can represent the true situation of the scanning, and the quality of the point cloud, such as the density, integrity or distortion degree of the point cloud, directly affects the accuracy of the environmental perception. If the quality of the point cloud is poor (such as sparse and blocked), the system can adjust to reduce the dependence on such data of the laser radar.

[0057] In a specific scheme, the first scum thickness and the second scum thickness are inputted in step S40, the corrected thickness is calculated according to the first confidence of the thickness measuring device and the second confidence of the laser radar, and one of the first scum thickness, the second scum thickness or the corrected thickness is selected as the working thickness according to the range of the first confidence and the second confidence, including:

[0058] In step S41, it is judged whether any one of the first confidence and the second confidence exceeds the first threshold value.

[0059] In step S42, if the first confidence exceeds the first threshold value, the first scum thickness is taken as the working thickness.

[0060] In step S43, if the second confidence exceeds the first threshold value, the second scum thickness is taken as the working thickness.

[0061] In step S44, if neither the first confidence nor the second confidence exceeds the first threshold value, the corrected thickness is taken as the working thickness.

[0062] It can be understood that, as known from the above, the first scum thickness is measured by the thickness measuring device (such as an ultrasonic / infrared sensor); the second scum thickness is measured by the laser radar; and the corrected thickness is a comprehensive value calculated in combination with the data of both (such as a weighted average). In step S42, it can be understood that if the confidence of the thickness measuring device is higher (exceeds a threshold value), the data (the first scum thickness) of the thickness measuring device is directly used. Conversely, in step S43, if the confidence of the laser radar is higher (exceeds a threshold value), the data (the second scum thickness) of the laser radar is directly used. That is, when neither of the two has a significant advantage, the corrected comprehensive value is used to improve the accuracy; the thickness measuring device (such as an ultrasonic sensor) and the laser radar may generate errors due to environmental interference (such as uneven scum surface, liquid splashing, and material reflectivity difference), and directly using the data of a single sensor may lead to an incorrect decision (such as misjudging the scum thickness, affecting the subsequent processing process).

[0063] In one scheme, the automatic scumming method of the electrolytic aluminum cell further comprises:

[0064] Step S401, taking the n most recently collected measurement values of the thickness measuring device, calculating the standard deviation of the n measurement values, and the signal-to-noise ratio of the n measurement values;

[0065] Step S402, inputting the standard deviation into a first quantization function to obtain a first confidence contribution value, wherein the first quantization function is monotonically decreasing with the increase of the standard deviation and always positive; inputting the signal-to-noise ratio into a second quantization function to obtain a second confidence contribution value, wherein the second quantization function is a tangent function;

[0066] Step S403, combining the first confidence contribution value and the second confidence contribution value to obtain the first confidence.

[0067] Taking the 100 most recently collected measurement values of the thickness measuring device (sampling frequency 10 Hz, i.e. 10 second window). The statistical quantities are calculated as follows: the first mean μ1 and the first standard deviation σ1, and the signal-to-noise ratio SNR. Confidence calculation: the first confidence contribution value C1 = α1·e -β1σ1 + α2·tanh(β2·SNR); wherein a1, a2, β1, β2 are adjustable coefficients, the sum of a1 and a2 is 1, and an example is as follows: the electrolytic cell 5 is measured at a temperature of 30°C, and the data within 10 seconds are μ = 5.12 cm, σ = 0.08 cm, and SNR = 25 dB. The obtained first confidence C1 = 0.5×e -2×0.08 + 0.5×tanh(0.1×25) = 0.914.

[0068] It should be noted that in the above formula for dynamically evaluating the first confidence level, the hyperbolic tangent function (tanh) is used to quantify the contribution of the signal-to-noise ratio (SNR) to the confidence level, mainly based on its mathematical properties and the matching of actual needs. Since tanh(x) is a continuous and derivable function, its output presents a smooth S-shaped curve with the change of input x. This property makes the confidence level gradually increase with the improvement of SNR, rather than suddenly jump. If f(SNR) = β2·SNR is directly used, when SNR is low, the confidence level may be negative (unreasonable); when SNR is high, the confidence level may exceed 1 (no upper limit), resulting in numerical instability. However, by using the above formula, when SNR is high, i.e. when SNR >> 1 / β2, tanh(β2·SNR) ≈ 1, indicating that the signal quality is extremely high, and the confidence level is close to the upper limit, avoiding the excessive amplification of the influence of small SNR improvement. Conversely, in a low SNR scenario, i.e. when SNR << 1 / β2, tanh(β2·SNR) ≈ β2·SNR, maintaining a linear response to ensure that the negative impact of low-quality signals on the confidence level is reasonably quantified.

[0069] In another alternative, the automatic slag recovery method of the electrolytic aluminum tank further comprises:

[0070] Step S404, dividing the liquid surface area point cloud into equal-sized grids with a plane, counting the number of points in each grid and calculating the second standard deviation;

[0071] Step S405, marking each grid with a point number less than the point number threshold as a low-density area. If at least one of the following conditions is met: the number of low-density area grids is greater than the number threshold or the second standard deviation is greater than the standard deviation threshold, then set the second confidence level to be lower than a pre-set confidence value, which is less than 1 minus the first threshold; if at least one of the following conditions is not met: the number of low-density area grids is greater than the number threshold or the second standard deviation is greater than the standard deviation threshold, then obtain the second confidence level based on the ideal standard deviation and the standard deviation under the pre-set scenario.

[0072] Specifically, assuming that the liquid surface area point cloud set P = {p1, p2,..., pN}, it can be directly divided into several equal-sized grids in the x-y plane by dividing the plane range; for each point pi, calculate its projection coordinates (xi', yi') on the plane and assign a grid index; in each grid, count the number of points to calculate the second mean μ2 and the second standard deviation σ2; by pre-setting a point number threshold T num , if μ2 in a certain grid is less than the point number threshold, mark it as a low-density area, and count the number of grids belonging to the low-density area N low ; if any of the following conditions is met: N low > T num or σ2 > T σIn this case, the second confidence is directly set to a fixed preset confidence value, which is a value less than the first threshold, so that the second confidence cannot exceed the first threshold.

[0073] If at least one of the number of low-density regions is greater than the number threshold or the second standard deviation is greater than the standard deviation threshold, the second confidence is obtained based on the ideal standard deviation and the standard deviation under the preset scene: Wherein, γ1 is an adjustable coefficient, C2 is the second confidence, σplane is the ideal standard deviation obtained under the ideal measurement; an exemplary calculation is: confidence: Clidar = 2·(1-0.3 / 0.5) = 0.8.

[0074] Then, a modified thickness is obtained according to the weighted fusion, which can be obtained by using the weighted average method, and in one embodiment, as follows: Tfinal = C1·T1 + C2·T2; wherein final is the modified thickness, T1 is the first scum thickness, and T2 is the second scum thickness.

[0075] In one embodiment of the present application, in step S30, the target electrolytic cell is scanned and the liquid surface region point cloud is segmented, and the second scum thickness is extracted from the liquid surface region point cloud, including:

[0076] Step S31, scanning the target electrolytic cell to obtain a point cloud set;

[0077] Step S32, extracting a local point cloud under a preset height range from the point cloud set, and performing plane fitting on the local point cloud to obtain a reference segmentation plane;

[0078] Step S33, segmenting the liquid surface region point cloud with the reference segmentation plane as a boundary reference, and obtaining the second scum thickness by calculating the distance from the liquid surface region point cloud to the reference segmentation plane.

[0079] It can be understood that the point cloud set includes all point clouds at the target electrolytic cell, and according to the electrolytic cell design parameters (such as the expected liquid surface height range under prior experience), the local point cloud in the preset height range (for example, ±10cm above and below the liquid surface) is extracted from the global point cloud; the reference segmentation plane is fitted using RANSAC (random sample consensus) or least squares method, which represents the ideal liquid surface position (the liquid surface without scum). The reference segmentation plane is a boundary, which divides the point cloud into two parts, i.e., points above the reference plane (scum surface points) and points below the reference plane (electrolyte or electrolytic cell bottom points, which may not be directly used in this step). For each point in the liquid surface region point cloud, calculate its vertical distance to the reference plane, and sort all points by vertical distance size to select the average value of the top 10% as the second scum thickness.

[0080] Further, in step S50, a slagging movement path is calculated according to the position information of the target electrolytic cell and the working thickness, and the crane and the slagging shovel of the crane are controlled to perform slagging according to the slagging movement path, including:

[0081] Step S51, according to the position information, the initial position of the slagging shovel is located;

[0082] Step S52, according to the working thickness, the target height is obtained, the slagging shovel is controlled to descend from the initial position, and during the descent of the slagging shovel, the current height and the target height are used to adjust the speed of the crane by PID;

[0083] Wherein, when the slagging shovel is performing slagging, the control strength of the slagging shovel is mapped according to the working thickness.

[0084] Specifically, the position information of the electrolytic cell (such as the edge coordinates of the cell body, the center point coordinates) is spatially aligned with the mechanical coordinate system of the crane (such as the track coordinates, the hook height), and a unified reference frame is established; according to the pre-set slagging starting point, the initial coordinates of the slagging shovel in the three-dimensional space are calculated. According to the working thickness, the insertion depth of the slagging shovel is determined, which can be calculated according to the calculation result, for example, when the working thickness is U1, the target height h1 can be obtained, wherein the target height h1 refers to the node height at which the slagging shovel stops on the liquid surface to start the slagging work. It can be simply calculated according to the liquid surface reference plane height, the working thickness (scum thickness) and the safety margin. The core logic is to ensure that the slagging shovel is inserted into the scum layer with accurate depth, while avoiding excessive pressure to contact the liquid surface.

[0085] The slagging shovel is controlled to descend, the current height of the slagging shovel is collected in real time, the difference between the target height is input into the PID controller to output the lowering speed of the crane hook, so that the speed gradually decreases when approaching the liquid surface to reduce the vibration and improve the accuracy.

[0086] And when the slagging shovel contacts the scum, the control strength (such as hydraulic pressure or motor torque) of the shovel is dynamically adjusted according to the working thickness, and a nonlinear relationship between the working thickness and the control strength is established to reduce the strength when the scum is thin to prevent the scum from falling back to the liquid surface, and to increase the strength when the scum is thick to ensure that the shovel penetrates and completely lifts the scum.

[0087] For example, Fig. 3 In the second aspect of the present application, the present application also provides an automatic slagging system 100 of an electrolytic aluminum cell, including a crane 10, a laser radar 20, a thickness measuring device 30 and a controller 40; the laser radar 20 is connected to the crane 10; the thickness measuring device 30 is connected to the crane 10; the controller 40 is electrically connected to the crane 10, the laser radar 20 and the thickness measuring device 30, and is configured to execute the above-mentioned automatic slagging method.

[0088] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program. When the computer program is run by a processor, the steps of the automatic slag salvaging method of the aluminum electrolysis cell in the above method embodiment are executed. The storage medium can be a volatile or non-volatile computer readable storage medium.

[0089] The embodiment of the present application further provides a computer program product, which carries a program code. The instructions included in the program code can be used to execute the steps of the workpiece positioning method of the multi-stage model in the above method embodiment. For details, refer to the above method embodiment, which will not be repeated here.

[0090] The computer program product can be specifically implemented by means of hardware, software or a combination thereof. In an optional embodiment, the computer program product is specifically embodied as a computer storage medium. In another optional embodiment, the computer program product is specifically embodied as a software product, such as a software development kit (SDK) and the like.

[0091] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system and device can refer to the corresponding process in the above method embodiment, which will not be repeated here. In the several embodiments provided by the present application, it should be understood that the disclosed system, device and method can be implemented by other ways. The above-described device embodiments are only schematic, for example, the division of the units is only a logical function division, and there can be another division manner in actual implementation, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual coupling or direct coupling or communication connection between the units can be indirect coupling or communication connection through some communication interfaces, devices or units, and can be electrical, mechanical or other forms.

[0092] The units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. According to actual needs, some or all of the units can be selected to achieve the purpose of the embodiment.

[0093] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit.

[0094] If the functions are realized in the form of software function units and sold or used as independent products, they can be stored in a nonvolatile computer readable storage medium executable by a processor. Based on this understanding, the technical solutions of the present application or the part of the prior art that essentially contributes or the part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes instructions for making a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0095] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present application, which are used to illustrate the technical solutions of the present application, but not to limit them. The protection scope of the present application is not limited thereto. Although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can make modifications or easily think of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed by the present application, or make equivalent replacements to some of the technical features. These modifications, changes or replacements do not make the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An automatic slag removal method for an electrolytic aluminum cell, characterized in that: include: Control the LiDAR to scan the electrolytic cell globally and locate the electrolytic cell through feature extraction; generating a cyclic trajectory based on the position information, controlling the overhead crane to move along the cyclic trajectory, and simultaneously obtaining first scum thicknesses in a plurality of electrolytic cells using a thickness measuring device on the overhead crane, and performing a threshold determination on the first scum thickness using a preset thickness threshold to obtain a target electrolytic cell from the plurality of electrolytic cells; Scanning the target electrolytic cell and segmenting the liquid surface area point cloud, and extracting the second scum thickness through the liquid surface area point cloud; Inputting the first scum thickness and the second scum thickness, respectively obtaining corrected thicknesses based on a first confidence level of a thickness measuring device and a second confidence level of a laser radar, and selecting the first scum thickness, the second scum thickness, or the corrected thickness as the working thickness based on the range of the first confidence level and the second confidence level; The slag removal movement path is calculated according to the position information and working thickness of the target electrolytic cell, and the movement of the overhead crane and the slag removal shovel of the overhead crane are controlled according to the slag removal movement path to remove the slag.

2. The automatic slag removal method for an electrolytic aluminum cell according to claim 1, characterized in that: The first confidence level is dynamically evaluated based on a standard deviation and a signal-to-noise ratio of the measurements, and the second confidence level is dynamically evaluated based on the quality of the scanned point cloud.

3. The automatic slag removal method for an electrolytic aluminum cell according to claim 2, characterized in that: The first scum thickness and the second scum thickness are input, and a corrected thickness is obtained according to a first confidence level of a thickness measuring device and a second confidence level of a laser radar, respectively, and one of the first scum thickness, the second scum thickness, or the corrected thickness is selected as the working thickness according to a range of the first confidence level and the second confidence level, including: Determining whether either the normalized first confidence level or the second confidence level exceeds a first threshold; If the first confidence level exceeds the first threshold, the first scum thickness is used as the working thickness; If the second confidence level exceeds the first threshold, the second scum thickness is used as the working thickness; If both the first confidence level and the second confidence level do not exceed a first threshold, the corrected thickness is used as the working thickness.

4. The automatic slag removal method for an electrolytic aluminum cell according to claim 3, characterized in that: The automatic slag removal method for the electrolytic aluminum cell also includes: Take the n most recent measurement values ​​collected by the thickness measuring device, calculate the first standard deviation of the n measurement values, and the signal-to-noise ratio of the n measurement values; Inputting the first standard deviation into a first quantization function to obtain a first confidence contribution value, wherein the first quantization function is monotonically decreasing with the increase of the standard deviation and is always a positive value; inputting the signal-to-noise ratio into a second quantization function to obtain a second confidence contribution value, wherein the second quantization function is a tangent function; The first confidence contribution value and the second confidence contribution value are combined to obtain a first confidence.

5. The automatic slag removal method for an electrolytic aluminum cell according to claim 3, characterized in that: The automatic slag removal method for the electrolytic aluminum cell also includes: Dividing the liquid surface area point cloud into grids of equal size using a plane projection, counting the number of points in each grid and calculating the second standard deviation of the number of points; Marking each grid with a number of points less than a point number threshold as a low-density area, and if at least one of the number of the low-density areas is greater than a number threshold or the second standard deviation is greater than a standard deviation threshold, setting the second confidence level to be lower than a preset confidence level, wherein the preset confidence level is less than the first threshold; If at least one of the number of the low-density areas being greater than the number threshold or the second standard deviation being greater than the standard deviation threshold does not exist, a second confidence level is obtained based on an ideal standard deviation in a preset scenario and the standard deviation.

6. The automatic slag removal method for an electrolytic aluminum cell according to claim 1, characterized in that: The scanning of the target electrolytic cell and segmenting the liquid surface area point cloud, and extracting the second scum thickness through the liquid surface area point cloud, comprises: Scanning the target electrolytic cell to obtain a point cloud set; Extracting a local point cloud within a preset height range from the point cloud set, and performing plane fitting in the local point cloud to obtain a reference segmentation plane; The liquid surface area point cloud is segmented with the reference segmentation plane as a boundary reference, and the second scum thickness is obtained by calculating the distance from the liquid surface area point cloud to the reference segmentation plane.

7. The automatic slag removal method for an electrolytic aluminum cell according to any one of claims 1 to 5, characterized in that: The calculating of the slag removal movement path according to the position information and the working thickness of the target electrolytic cell, and controlling the overhead crane and the slag removal shovel of the overhead crane to remove the slag according to the slag removal movement path, includes: locating an initial position of the slag shovel according to the position information; Obtain the target height based on the working thickness, control the slag shovel to descend from the initial position, and use PID to adjust the overhead crane speed according to the current height and target height during the slag descent process; When the slag shovel is scooping slag, the control force of the slag shovel is mapped according to the working thickness.

8. The automatic slag removal method for an electrolytic aluminum cell according to any one of claims 1 to 5, characterized in that: Also includes: After the slag is removed, the gravity sensor is controlled to measure the amount of floating slag, and the overhead crane is dispatched to remove the slag according to the preset slag box position; Controlling the overhead crane to return to the target electrolytic cell, controlling the thickness measuring device to measure the first slag thickness again, and controlling the laser radar to scan again and calculate the second slag thickness; A threshold judgment is performed on both the first slag thickness or the second slag thickness. If both are less than the thickness threshold, the corresponding target electrolytic cell slag is completed; if either one is greater than the thickness threshold, the slag scoop movement path is continued to be generated and the overhead crane and the slag scoop of the overhead crane are controlled to scoop the slag.

9. An automatic slag removal system for an electrolytic aluminum cell, characterized in that: include: Overhead crane; a laser radar connected to the overhead crane; a thickness measuring device connected to the overhead traveling crane; A controller is electrically connected to the overhead crane, the laser radar, and the thickness measuring device, and is configured to execute the automatic slag dredging method according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, executes the steps of the automatic slag scooping method for an electrolytic aluminum cell as claimed in any one of claims 1 to 8.

Images