Grinding control method and system of roll grinder

Abstract

The invention relates to the technical field of roll grinders, and discloses a grinding control method and system for a roll grinder, and the method comprises the steps: deploying a pressure sensor, collecting pressure data and position coordinates, and building an initial data set. Abnormal data are identified and removed by detecting a time-pressure curve, missing values are complemented by utilizing curve fitting, and a corrected data set is formed. Workpiece boundaries are recognized based on pressure distribution, and after effective data are screened, a weighted average pressure value is calculated in the vertical direction of the grinding wheel. And dynamically adjusting the normal force of the grinding wheel by adopting a proportional control algorithm to enable the pressure to approach a target value. Abnormal data can be monitored, equipment health state evaluation is supported, the workpiece surface burning risk is reduced through pressure feedback compensation, and the size precision is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of roll grinders, and in particular, to a grinding control method and system for a roll grinder. Background Art

[0002] Roll grinding is a key process to ensure the accuracy of rolled steel and non-ferrous metal rolling products, and its processing quality directly affects core indicators such as the thickness uniformity and surface finish of the sheet. Traditional roll grinders generally adopt an open-loop control method based on preset feed parameters, relying on manual experience to set parameters such as grinding depth and feed speed, resulting in the technical problem that the dynamic fluctuation of the normal force is difficult to control. Factors such as wheel wear, uneven workpiece material, and uneven distribution of machining allowance will cause changes in contact pressure. Measured data shows that the pressure fluctuation in the edge area can reach ±30% of the mean value, far exceeding the ±10% deviation allowed by the ISO468 standard, which in turn leads to defects such as surface burn and dimensional deviation. The traditional process relies on off-line coordinate measuring machines and cannot track the change in the diameter of the workpiece during the grinding process (such as the material removal per revolution is 0.01 - 0.05 mm), resulting in a mismatch between the feed parameters and the actual machining state. At the same time, the fixed feed strategy needs to reduce the feed speed to less than 0.1 mm / min in the fine grinding stage, reducing the processing efficiency by more than 60%, and cannot compensate for the change in contact stiffness caused by wheel wear (when the wear increases the contact area by 20%, the normal force spontaneously rises by 15% - 25%).

[0003] Although existing improvement schemes introduce a single-axis force sensor, it can only monitor the total grinding force and cannot reflect the spatial difference in pressure distribution. There is still a dimensional deviation of ±20 μm in complex surface grinding. Industry statistics show that the defect rate of surface burn in the traditional process is as high as 3% - 5%, increasing the repair cost of a single roll. Summary of the Invention

[0004] In view of this, the present invention proposes a grinding control method and system for a roll grinder, aiming to solve the problems in the current technology that the roll grinder lacks a force distribution perception and compensation mechanism, resulting in unstable processing quality caused by normal force fluctuations, and it is difficult to balance efficiency and accuracy.

[0005] A grinding control method for a roll grinder proposed by the present invention includes: A plurality of pressure sensors with known position coordinates are arranged above the workpiece to be ground. When the pressure sensors contact the workpiece to be ground, the pressure data and position coordinates of each pressure sensor are periodically collected, and an initial data set is established according to the pressure data, position coordinates, and corresponding detection time obtained each time; Judge the abnormal data in the initial data set, record the position coordinates corresponding to the abnormal data, eliminate the abnormal data, complete the eliminated abnormal data, and use the current initial data set as the corrected data set; obtain the pressure data in the corrected data set, identify the size and position information of the workpiece to be ground according to the pressure data, select the pressure sensors under the workpiece to be ground, and obtain the data in the corrected data set according to the selected pressure sensors, which is recorded as the effective data set; Obtain the running direction of the grinding wheel, select the pressure data of the pressure sensors on the same straight line perpendicular to the running direction according to the position coordinates, and calculate the average pressure value; Adjust the normal force applied by the grinding wheel to the workpiece to be ground according to the average pressure value.

[0006] Further, when periodically collecting the pressure data and position coordinates of each pressure sensor, it further includes: Count the occurrence times of abnormal data in each pressure sensor, compare the occurrence times with a preset number threshold, and when the occurrence times are greater than or equal to the number threshold, send an alarm signal.

[0007] Further, the abnormal data judgment includes: Establish a coordinate system with the detection time as the x-axis and the pressure data as the y-axis, obtain the detection time and pressure data in the initial data set, and use the detection time and the corresponding pressure data as data points. Obtain the detection time-pressure data curve of the initial data set by fitting. Calculate the y-axis distance from each data point to the curve at the same detection time, calculate the average value and variance of the y-axis distance. When the y-axis distance is greater than the average value of the y-axis distance + variance or less than the average value of the y-axis distance - variance, determine that the pressure data of the current data point is abnormal data.

[0008] Further, when completing the eliminated abnormal data, obtain the detection time of the abnormal data, and replace the abnormal data with the pressure data obtained by substituting the detection time into the detection time-pressure data curve.

[0009] Further, when identifying the size and position information of the workpiece to be ground according to the pressure data, it includes: Use the random forest model to divide the pressure data and corresponding position coordinates in the corrected data set into training set, validation set and test set for training. First, label the pressure data as below the workpiece to be ground and not below the workpiece to be ground respectively, obtain the pressure data below the workpiece to be ground, determine the boundary of the workpiece to be ground according to the position coordinates corresponding to the pressure data, so as to determine the position information, and calculate the size of the workpiece to be ground.

[0010] Further, when selecting the pressure data of the pressure sensors on the same straight line perpendicular to the running direction according to the position coordinates and calculating the average pressure value, it includes: ; wherein, is the average pressure value on the j-th straight line, is the number of pressure sensors on the same straight line, is the pressure data weight coefficient of the i-th pressure sensor on the same straight line, is the pressure data of the i-th pressure sensor on the same straight line.

[0011] Further, the pressure data weight coefficient is obtained through the following relationship: ; wherein, is the position coordinate of the i-th pressure sensor on the same straight line, is a non-zero constant, = 10 -6 , is the geometric center of the to-be-ground surface of the workpiece to be ground.

[0012] Further, when adjusting the normal force applied by the grinding wheel to the workpiece to be ground according to the average pressure value, it includes: Calculating an adjustment coefficient by calculating the average pressure value and a preset standard pressure value, and adjusting the normal force according to the adjustment coefficient:

[0013] wherein, is the adjusted normal force, is the standard pressure value, is the normal force on the j-th straight line.

[0014] The present invention also provides a grinding control system for a roll grinder, including: A data acquisition module, configured to set a plurality of pressure sensors with known position coordinates above the workpiece to be ground. When the pressure sensors contact the workpiece to be ground, the pressure data and position coordinates of each pressure sensor are periodically acquired, and an initial data set is established according to the pressure data, position coordinates and corresponding detection time obtained each time; A data processing module, configured to judge abnormal data in the initial data set, record the position coordinates corresponding to the abnormal data, eliminate the abnormal data, complete the eliminated abnormal data, and use the current initial data set as a corrected data set; obtain the pressure data in the corrected data set, identify the size and position information of the workpiece to be ground according to the pressure data, select the pressure sensors located below the workpiece to be ground, and obtain the data in the corrected data set according to the selected pressure sensors, and record it as an effective data set; An adjustment control module, configured to obtain the running direction of the grinding wheel, select the pressure data of the pressure sensors on the same straight line perpendicular to the running direction according to the position coordinates, and calculate the average pressure value; Adjust the normal force applied by the grinding wheel to the workpiece to be ground according to the average pressure value.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: By arranging a plurality of pressure sensors above the roll, the present invention can collect and monitor the normal force distribution during the grinding process. Compared with the traditional single-axis force sensor, this solution can reflect the spatial pressure difference on the roll surface and effectively identify local pressure abnormalities. By adjusting the normal force, the pressure fluctuation in the edge area is reduced, ensuring uniform grinding force distribution, thereby significantly reducing the risk of surface burn and dimensional tolerance. In the traditional process, due to the inability to monitor pressure changes, processing errors are likely to occur, while the present invention can control the normal force fluctuation within a reasonable range and improve the processing accuracy.

[0016] The present invention introduces an anomaly detection algorithm based on time and pressure data curves, which can accurately identify abnormal data points caused by sensor errors or jitter noise and automatically eliminate them. For the data after elimination, the system automatically completes the data through curve fitting, thus ensuring the continuity and accuracy of the data. Compared with the traditional process, this solution can continuously monitor and repair abnormal data during the grinding process, significantly improving the stability and reliability of the grinding data.

[0017] The present invention uses a random forest model to intelligently identify the pressure sensor data, which can automatically distinguish the area to be ground from the non-grinding area and accurately determine the boundary and size information of the roll. In the traditional process, it usually relies on manual experience or offline detection, with low positioning accuracy and easy to cause dimensional deviation. The present invention automatically identifies the boundary and size through a machine learning model, greatly improving the grinding accuracy and ensuring processing consistency.

[0018] The present invention calculates the average pressure value based on the pressure distribution data, compares it with the standard pressure value, automatically generates an adjustment coefficient, and realizes the dynamic compensation of the normal force of the grinding wheel. When the contact stiffness changes due to the wear of the grinding wheel, the system can adaptively adjust the normal force to prevent the fluctuation of the machining quality. In the traditional process, the wear of the grinding wheel will cause a change in the contact area, resulting in an increase in the normal force, thereby affecting the machining accuracy. The present invention can automatically compensate for the change in the normal force, ensure the consistency of the grinding effect, and guarantee the stability of the machining quality.

[0019] In the fine grinding stage of the traditional grinding process, the feed speed often needs to be reduced to an extremely low level to prevent burns and accuracy errors, but this will significantly reduce the machining efficiency. Through pressure monitoring and dynamic normal force compensation, the present invention can maintain a high feed speed while ensuring the machining quality, effectively improving the machining efficiency. Compared with the traditional process, under the premise of ensuring accuracy, the machining time is significantly shortened, and the production efficiency is improved.

[0020] The present invention has an intelligent alarm function. When the number of abnormal data of the pressure sensor exceeds the preset threshold, the system will automatically send an alarm signal, effectively warning of sensor failures or abnormal grinding conditions. This mechanism can detect and handle equipment abnormalities in a timely manner, prevent the decline of machining quality caused by equipment problems, and thus significantly improve the stability and safety of the system.

[0021] It can be understood that the above-mentioned grinding control method and system of a roll grinder have the same beneficial effects and will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] By reading the following detailed description of the preferred embodiments, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present invention. Moreover, throughout the drawings, the same reference numerals are used to represent the same components. In the drawings: Figure 1 It is a schematic diagram of the working scenario of the roll grinder provided by the embodiment of the present invention.

[0023] Figure 2 It is a flowchart of a grinding control method for a roll grinder provided by the embodiment of the present invention.

[0024] Figure 3 It is a functional framework diagram of a grinding control system for a roll grinder provided by the embodiment of the present invention.

[0025] Among them, 1. Grinding wheel; 2. Roll; 3. Pressure sensor; 4. Sensor fixing plate; 5. Fixed bracket; 6. Grinding wheel fixing rod. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0026] The exemplary embodiments disclosed in the present application will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, in the absence of conflict, the embodiments of the present invention and the features described in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0027] See also Figure 1-2 As shown, an embodiment of the present invention provides a grinding control method for a roll grinder, comprising: S1: a plurality of pressure sensors with known position coordinates are arranged above the workpiece to be ground. When the pressure sensors are in contact with the workpiece to be ground, the pressure data and position coordinates of each pressure sensor are periodically collected, and an initial data set is established according to the pressure data, position coordinates and corresponding detection time obtained each time; S2: judge the abnormal data of the initial data set, record the position coordinates corresponding to the abnormal data and remove the abnormal data, complete the removed abnormal data, and use the current initial data set as the corrected data set; obtain the pressure data in the corrected data set, identify the size and position information of the workpiece to be ground according to the pressure data, select the pressure sensor under the workpiece to be ground, obtain the data in the corrected data set according to the selected pressure sensor, and record it as a valid data set; S3: Obtain the running direction of the grinding wheel, select the pressure data of the pressure sensor on the same straight line perpendicular to the running direction according to the position coordinates, and calculate the average pressure value; The normal force applied by the grinding wheel to the workpiece to be ground is adjusted according to the average pressure value.

[0028] It should be noted that, see Figure 1 As shown, when the grinding wheel 1 and the roller 2 (i.e., the workpiece to be ground) stop rotating, the fixed bracket 5 pushes the sensor fixing plate 4, so that the pressure sensor 3 arranged on the sensor fixing plate 4 comes into contact with the surface to be ground of the roller 2, and then the reading of the pressure sensor 3 is collected; when the grinding wheel fixing rod 6 pushes the grinding wheel 1 to contact with the roller 2, and they start to rotate respectively, grinding starts.

[0029] When collecting data, the system will periodically record the pressure value and corresponding coordinate position of each sensor, and mark the detection time to form an initial data set. In the data processing stage, the system automatically identifies and removes abnormal data in the data set through anomaly detection algorithms to ensure data accuracy. After removing abnormal data, the system will automatically complete the missing data based on the time and pressure data curve to ensure data continuity and stability.

[0030] During the grinding process, the system can analyze the corrected data set through a random forest model to automatically identify the size and position information of the workpiece to be ground. By annotating and training the sensor data, the model accurately judges the sensor distribution under the workpiece to be ground, thereby realizing the automatic identification and size measurement of the grinding workpiece boundary and ensuring the machining positioning accuracy.

[0031] In the normal force adjustment stage, the system selects the pressure sensor data on the same straight line perpendicular to the running direction of the grinding wheel, calculates the average pressure value on this straight line, compares this average pressure value with the preset standard pressure value to generate an adjustment coefficient, and dynamically adjusts the normal force applied by the grinding wheel accordingly. When the pressure distribution fluctuates or the contact stiffness changes due to grinding wheel wear, the system can automatically compensate the normal force to ensure the consistency of grinding accuracy and surface quality.

[0032] In addition, this embodiment also introduces an intelligent alarm mechanism. During the grinding process, the system continuously monitors the abnormal conditions of the sensor data. When the number of occurrences of abnormal data of a certain sensor exceeds the preset threshold, the system automatically sends an alarm signal to indicate that there may be a sensor failure or grinding abnormality, which is convenient for timely maintenance and handling, thereby further improving the safety and stability of the equipment operation.

[0033] In some embodiments of the present application, when periodically collecting the pressure data and position coordinates of each pressure sensor, it further includes: Count the number of occurrences of abnormal data for each pressure sensor, compare the number of occurrences with the preset number threshold, and send an alarm signal when the number of occurrences is greater than or equal to the number threshold.

[0034] Specifically, when periodically collecting the pressure data and position coordinates of each pressure sensor, it further includes the statistics of abnormal data and the alarm mechanism. Specifically, after each data collection, the system detects the pressure values of each pressure sensor and judges whether there is abnormal data. Once abnormal data is detected, the system records the number of occurrences of this abnormal data and compares it with the preset number threshold.

[0035] When the number of occurrences of abnormal data of a certain sensor reaches or exceeds the preset number threshold, the system automatically sends an alarm signal to indicate that there may be a sensor failure or grinding abnormality. This intelligent alarm mechanism can detect abnormalities in a timely manner, effectively prevent the decline of grinding accuracy or processing defects caused by sensor failure or data abnormality. At the same time, this mechanism can also remind the operator to repair or maintain the equipment to ensure the stability and safety of the grinding process.

[0036] In addition, through the cumulative statistics of abnormal data and comparison with the threshold value, occasional noise or error signals can be filtered out, and an alarm is triggered only when sensor abnormalities occur frequently, thereby improving the accuracy and reliability of the alarm. This mechanism helps to ensure the data quality and processing consistency during the grinding process.

[0037] In some embodiments of the present application, the abnormal data determination includes: Taking the detection time as the x-axis and the pressure data as the y-axis to establish a coordinate system, obtaining the detection time and pressure data in the initial data set, and using the detection time and the corresponding pressure data as data points. By fitting, the detection time-pressure data curve of the initial data set is obtained. At the same detection time, calculate the y-axis distance from each data point to the curve, calculate the average value and variance of the y-axis distance. When the y-axis distance is greater than the average value of the y-axis distance + variance or less than the average value of the y-axis distance - variance, determine that the pressure data of the current data point is abnormal data.

[0038] It should be noted that in the process of abnormal data determination in this embodiment, by constructing a time-pressure coordinate system with the detection time and pressure data, and using the curve fitting method to analyze the data, abnormal data points can be effectively identified. Specifically, the system generates a time-pressure data curve through fitting, which is used as a reference benchmark for normal pressure changes. During the detection process, the system calculates the distance from each data point to the fitting curve in the vertical axis direction, and uses the average value and variance of this distance as the judgment basis. When the distance of a certain data point exceeds the average value plus the variance or is lower than the average value minus the variance, it can be determined that this data point is abnormal data.

[0039] The beneficial effects of this abnormal judgment method are as follows: Accurately identify abnormal data: By adopting the curve fitting and distance deviation judgment methods, abnormal data points caused by sensor errors, signal jitter or instantaneous interference can be effectively filtered out, ensuring the accuracy of the data. Compared with the traditional fixed threshold judgment method, this method has stronger adaptability and robustness.

[0040] Improve data stability: During the grinding process, eliminating abnormal data and completing data filling can ensure the continuity and stability of the grinding data, and prevent the deviation of the normal force control caused by individual abnormal data.

[0041] Enhance grinding accuracy and consistency: By eliminating abnormal data, the system can adjust based on more accurate data when calculating the normal force compensation, ensuring that the normal force applied by the grinding wheel is consistent with the expected value, thereby improving the grinding accuracy and surface quality.

[0042] Reducing false alarms and misjudgments: Through curve fitting and variance judgment, the system can effectively distinguish normal fluctuations from abnormal data, avoid misjudging data fluctuations or noise as abnormal, and improve the accuracy and stability of anomaly detection.

[0043] In some embodiments of the present application, when complementing the eliminated abnormal data, the detection time of the abnormal data is obtained, and the pressure data obtained by substituting this detection time into the detection time-pressure data curve is used to replace the abnormal data.

[0044] In some embodiments of the present application, when identifying the size and position information of the workpiece to be ground based on the pressure data, it includes: Using a random forest model to divide the pressure data and the corresponding position coordinates in the corrected dataset into a training set, a validation set, and a test set for training. First, the pressure data is respectively labeled as below the workpiece to be ground and not below the workpiece to be ground, the pressure data below the workpiece to be ground is obtained, and the boundary of the workpiece to be ground is determined based on the position coordinates corresponding to the pressure data, thereby determining the position information and calculating the size of the workpiece to be ground.

[0045] It should be noted that when identifying the size and position information of the workpiece to be ground in this embodiment, by introducing a random forest model for data training and classification, the accuracy and stability of the grinding workpiece boundary recognition can be significantly improved. Specifically, the system first divides the pressure data and the corresponding position coordinates in the corrected dataset into a training set, a validation set, and a test set, and uses the random forest model for training. The model automatically labels the sensor pressure data as the categories of below the workpiece to be ground and not below the workpiece to be ground, thereby realizing the distinction between the grinding area and the non-grinding area.

[0046] Through automatic classification by the machine learning model, it can accurately distinguish the workpiece to be ground from the surrounding non-processing areas, and automatically calculate the boundary and position information of the grinding workpiece. Compared with the traditional manual discrimination or the recognition method based on a fixed threshold, this method can dynamically adapt to different workpiece sizes and shapes, and has higher recognition accuracy. The random forest model can make judgments by voting multiple decision trees and comprehensively considering multi-dimensional features, thereby improving the accuracy of grinding boundary recognition. This method effectively reduces the boundary recognition error, ensures that the grinding process accurately covers the area to be processed, and prevents dimensional over-tolerance or surface defects caused by inaccurate boundary recognition. The random forest model can automatically learn the characteristics of different workpiece shapes and sizes during the training process, and has strong self-adaptability. Even if the workpiece deforms or shifts during the grinding process, the system can still accurately identify the workpiece boundary and ensure the grinding accuracy. By automatically identifying the grinding area through machine learning, without relying on manual judgment, it effectively reduces human misjudgment and operation deviation, and further improves the automation and intelligence level of the grinding process.

[0047] In some embodiments of the present application, when selecting the pressure data of the pressure sensors on the same straight line perpendicular to the running direction according to the position coordinates and calculating the average pressure value, it includes: ; wherein, is the average pressure value on the j-th straight line, is the number of pressure sensors on the same straight line, is the pressure data weight coefficient of the i-th pressure sensor on the same straight line, is the pressure data of the i-th pressure sensor on the same straight line.

[0048] In some embodiments of the present application, the pressure data weight coefficient is obtained through the following relationship: ; wherein, is the position coordinate of the i-th pressure sensor on the same straight line, is a non-zero constant, = 10 -6 , is the geometric center of the surface to be ground of the workpiece to be ground.

[0049] It should be noted that during the grinding process of the roll grinder, the pressures received by different positions of the workpiece to be ground are usually different. Generally speaking, the closer to the center position of the workpiece to be ground, the relatively more stable the pressure and the greater the impact on the overall grinding effect; while at the edge position, due to factors such as stress concentration and uneven material distribution, the pressure change may be more complex and the representativeness of the overall grinding effect is relatively weak. For example, the force on the center part of the roll is relatively uniform during grinding, which can better reflect the overall grinding state, while the edge part may be affected by external factors and the pressure fluctuates greatly. By introducing the distance between the current coordinate and the center coordinate to calculate the weight, the data collected by the sensors closer to the center can account for a larger proportion in the weight calculation, so as to more accurately reflect the actual pressure distribution of the workpiece to be ground. During the grinding process of the roll grinder, accurately grasping the pressure distribution of the workpiece to be ground is crucial for adjusting the normal force applied by the grinding wheel. By considering the distance between the current coordinate and the center coordinate to calculate the weight, the pressure conditions at different positions of the workpiece to be ground can be determined more accurately, so as to adjust the normal force of the grinding wheel according to the actual pressure distribution and achieve more precise grinding control. For example, when it is detected that the pressure at the center position is large, the normal force of the grinding wheel corresponding to this position can be appropriately reduced to avoid excessive grinding; while when the pressure at the edge position is small, the normal force can be appropriately increased to ensure the uniformity of grinding.

[0050] In some embodiments of the present application, when adjusting the normal force applied by the grinding wheel to the workpiece to be ground according to the average pressure value, it includes: The average pressure value is calculated with the preset standard pressure value to obtain the adjustment coefficient, and the normal force is adjusted according to the adjustment coefficient:

[0051] in, is the adjusted normal force, is the standard pressure value, is the normal force on the jth line.

[0052] It should be noted that by dynamically comparing the average pressure value with the standard pressure value, the system can generate an adjustment coefficient so that the normal force automatically adapts to the actual force of the workpiece during processing. When the grinding wheel wears or the workpiece material is uneven, causing the grinding contact stiffness to change, the system can adjust the normal force in real time to prevent surface burns caused by excessive normal force or dimensional errors caused by insufficient normal force. This adaptive compensation mechanism ensures balanced pressure during the grinding process, thereby significantly improving grinding accuracy and dimensional consistency.

[0053] See also Figure 3 As shown, this embodiment provides a grinding control system for a roll grinder, comprising: A data acquisition module is configured to set a plurality of pressure sensors with known position coordinates above the workpiece to be ground, and periodically collect the pressure data and position coordinates of each pressure sensor when the workpiece to be ground contacts the pressure sensor, and establish an initial data set according to the pressure data, position coordinates and corresponding detection time obtained each time; The data processing module is configured to judge abnormal data of the initial data set, record the position coordinates corresponding to the abnormal data and remove the abnormal data, complete the removed abnormal data, and use the current initial data set as the corrected data set; obtain pressure data in the corrected data set, identify the size and position information of the workpiece to be ground according to the pressure data, select a pressure sensor under the workpiece to be ground, obtain data in the corrected data set according to the selected pressure sensor, and record it as a valid data set; The adjustment control module is configured to obtain the running direction of the grinding wheel, select the pressure data of the pressure sensor on the same straight line perpendicular to the running direction according to the position coordinates, and calculate the average pressure value; The normal force applied by the grinding wheel to the workpiece to be ground is adjusted according to the average pressure value.

[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent substitutions can still be made to the specific implementation manners of the present invention, and any modification or equivalent substitution that does not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A grinding control method for a roll grinder, characterized in that, include: A plurality of pressure sensors with known position coordinates are arranged above the workpiece to be ground. When the pressure sensor contacts the pressure sensor through the workpiece to be ground, the pressure data and position coordinates of each pressure sensor are periodically collected, and an initial data set is established according to the pressure data, position coordinates and corresponding detection time obtained each time; Perform abnormal data judgment on the initial data set, record the position coordinates corresponding to the abnormal data and remove the abnormal data, complete the removed abnormal data, and use the current initial data set as a revised data set; Obtaining pressure data in the correction data set, identifying the size and position information of the workpiece to be ground according to the pressure data, selecting a pressure sensor located above the workpiece to be ground, obtaining data in the correction data set according to the selected pressure sensor, and recording it as a valid data set; Obtaining the running direction of the grinding wheel, selecting the pressure data of the pressure sensor on the same straight line perpendicular to the running direction according to the position coordinates, and calculating the average pressure value; The normal force applied by the grinding wheel to the workpiece to be ground is adjusted according to the average pressure value.

2. The grinding control method of the roll grinder according to claim 1, wherein, When the pressure data and position coordinates of each pressure sensor are periodically collected, the method further includes: The number of occurrences of abnormal data in each pressure sensor is counted, and the number of occurrences is compared with a preset number threshold. When the number of occurrences is greater than or equal to the number threshold, an alarm signal is issued.

3. The grinding control method of the roll grinder according to claim 1, wherein The abnormal data judgment includes: A coordinate system is established with the detection time as the x-axis and the pressure data as the y-axis, the detection time and pressure data in the initial data set are obtained, and the detection time and the corresponding pressure data are used as data points. The detection time-pressure data curve of the initial data set is obtained by fitting, and the y-axis distance of each data point to the curve is calculated at the same detection time, and the average value and variance of the y-axis distance are calculated. When the y-axis distance is greater than the average value + variance of the y-axis distance or less than the average value - variance of the y-axis distance, the pressure data of the current data point is determined to be abnormal data.

4. The grinding control method of the roll grinder according to claim 3, characterized in that, When the removed abnormal data is supplemented, the detection time of the abnormal data is obtained, and the pressure data obtained by substituting the detection time into the detection time-pressure data curve replaces the abnormal data.

5. The grinding control method of the roll grinder according to claim 1, characterized in that When the size and position information of the workpiece to be ground is identified according to the pressure data, it includes: A random forest model is used to divide the pressure data and the corresponding position coordinates in the corrected data set into a training set, a validation set, and a test set for training. The pressure data are first marked as below the part to be ground and not below the part to be ground, and the pressure data below the part to be ground is obtained. The boundary of the part to be ground is determined according to the position coordinates corresponding to the pressure data, thereby determining the position information and calculating the size of the part to be ground.

6. The grinding control method of the roll grinder according to claim 1, characterized in that, When the pressure data of the pressure sensor on the same straight line perpendicular to the running direction is selected according to the position coordinates, the average pressure value is calculated, including: ； Among them, is the average pressure value on the j-th straight line, is the number of pressure sensors on the same straight line, is the pressure data weight coefficient of the i-th pressure sensor on the same straight line, is the pressure data of the i-th pressure sensor on the same straight line.

7. The grinding control method of the roll grinder according to claim 1, characterized in that, The pressure data weight coefficient is obtained through the following relationship: ； Among them, is the position coordinate of the i-th pressure sensor on the same straight line, is a non-zero constant, = 10 -6 , is the geometric center of the surface to be ground of the workpiece to be ground.

8. The grinding control method of the roll grinder according to claim 1, characterized in that When the normal force applied by the grinding wheel to the workpiece to be ground is adjusted according to the average pressure value, it includes: The average pressure value and the preset standard pressure value are calculated to obtain an adjustment coefficient, and the normal force is adjusted according to the adjustment coefficient: ； Among them, is the adjusted normal force, is the standard pressure value, is the normal force on the j-th straight line.

9. A grinding control system for a roll grinder, applied to the grinding control system of the roll grinder according to any one of claims 1-8, characterized in that, include: A data acquisition module is configured to set a plurality of pressure sensors with known position coordinates above the workpiece to be ground, and periodically collect the pressure data and position coordinates of each pressure sensor when the workpiece to be ground contacts the pressure sensor, and establish an initial data set according to the pressure data, position coordinates and corresponding detection time obtained each time; A data processing module is configured to perform abnormal data judgment on the initial data set, record the position coordinates corresponding to the abnormal data and remove the abnormal data, complete the removed abnormal data, and use the current initial data set as a revised data set; Obtaining pressure data in the correction data set, identifying the size and position information of the workpiece to be ground according to the pressure data, selecting a pressure sensor located below the workpiece to be ground, obtaining data in the correction data set according to the selected pressure sensor, and recording it as a valid data set; The adjustment control module is configured to obtain the running direction of the grinding wheel, select the pressure data of the pressure sensor on the same straight line perpendicular to the running direction according to the position coordinates, and calculate the average pressure value; The normal force applied by the grinding wheel to the workpiece to be ground is adjusted according to the average pressure value.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the grinding control method for a roll grinder according to any one of claims 1 to 8.

Images