Data Processing Method and Device for Intelligent Polishing

Abstract

A data processing method and device for intelligent grinding, belonging to the technical field of electrical digital data processing, which includes: obtaining the weight sampling values of a number of sampling points at the driving end of the reverse driver transmitted by sampling; obtaining the trend coefficients of the weight sampling values at each time point in the sampling value queue of each sampling point, and thus obtaining the disturbed amplitude after calibration of the weight sampling values at each time point; obtaining the weights of the weight sampling values at each time point of each sampling point; and performing calibration on the weight sampling values of all sampling points at each time point through the weights of the weight sampling values at each time point of each sampling point to obtain the weight value of the grinding machine. It effectively avoids the defect in the prior art that taking the average of the weight values on all sampling points as the weight value of the grinding machine for intelligent grinding will weaken the accuracy of the weight value sampling of the grinding machine for intelligent grinding.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric digital data processing, and particularly relates to a data processing method and device for intelligent grinding. Background Art

[0002] Intelligent grinding is a grinding technology based on advanced automation technology and intelligent algorithms. It takes artificial intelligence as the core, is equipped with various sensors and actuators, and can simulate the actions and skills of manual operators to complete efficient and precise grinding tasks.

[0003] In practical applications, currently, intelligent grinding is often as mentioned in the prior art solution with the patent publication number "CN110842782B", which includes a weighing sensor, a position sensor, and a reverse driver connected to a controller. The weighing sensor is used to sample the weight value of the grinding machine and transmit the weight value to the controller. The weighing sensor is arranged at the driving end of the reverse driver. The position sensor is used to sample the position value of the grinding machine and transmit the position value to the controller. The controller is used to adjust the driving force output by the reverse driver according to the transmitted weight value and position value.

[0004] Specifically, to prevent the deviation caused by a single weighing sensor and for more accurate sampling of the weight value of the grinding machine, it is necessary to set several weighing sensors at several preset sampling points at the driving end of the reverse driver. In this way, the controller takes the average of the weight sampling values sampled by the weighing sensors at all sampling points at the same time point as the weight value of the grinding machine, and then adjusts the driving force output by the reverse driver according to the weight value and position value. However, because the interference amplitudes of different sampling points at the driving end of the reverse driver by the external magnetic field are different, and the typicality of the weight sampling values at different sampling points is different, taking the average of the weight values at all sampling points as the weight value of the grinding machine will reduce the accuracy of the weight value sampling of the grinding machine. Summary of the Invention

[0005] To solve the defects in the prior art, the present invention proposes a data processing device and method for intelligent grinding, effectively avoiding the defect that taking the average of the weight values at all sampling points as the weight value of the grinding machine for intelligent grinding in the prior art will reduce the accuracy of the weight value sampling of the grinding machine for intelligent grinding.

[0006] The present invention adopts the following technical solutions.

[0007] A data processing method for intelligent grinding, comprising:

[0008] The weighing sensor samples the weight sampling value of the grinding machine and transmits the weight sampling value to the controller. The position sensor samples the position value of the grinding machine and transmits the position value to the controller. The controller obtains the weight value of the grinding machine based on the weight sampling value transmitted by sampling, and then adjusts the driving force output by the reverse driver according to the weight value and position value of the grinding machine;

[0009] The method by which the controller obtains the weight value of the grinding machine based on the weight sampling value transmitted by sampling includes:

[0010] Step1, obtain the weight sampling values of several sampling points at the driving end of the reverse driver transmitted by sampling;

[0011] Step2, through the change trend of the weight sampling values in the sampling value queue of each sampling point, obtain the trend coefficient of the weight sampling values at each time point in the sampling value queue of each sampling point, and thus obtain the disturbance amplitude after calibration of the weight sampling values at each time point;

[0012] Step3, through the arrangement of the weight sampling values of all sampling points at each time point on the driving end of the reverse driver and the disturbance amplitude after calibration, obtain the weight of the weight sampling values at each time point of each sampling point;

[0013] Step4, through the weights of the weight sampling values at each time point of each sampling point, calibrate the weight sampling values of all sampling points at each time point to obtain the weight value of the grinding machine.

[0014] Further, in Step1, arrange the weight sampling values of all time points on each sampling point in the order of their sampling time points to form a queue, which is defined as the sampling value queue of each sampling point.

[0015] Further, in Step2, the specific method for obtaining the disturbance amplitude after calibration of the weight sampling values at each time point includes: performing regression processing on the sampling value queue of each sampling point to obtain the sampling value regression line of each sampling point, and obtaining the time domain coefficient of the sampling value queue of each sampling point through the difference in the time intervals between all adjacent peak-valley points in the sampling value regression line; through the trend coefficient and the time domain coefficient, obtain the sampling value disturbance amplitude at each time point of each sampling point; through the average of the similarities between the sampling value queues of each sampling point and those of other sampling points, calibrate the sampling value disturbance amplitude at each time point of each sampling point to obtain the disturbance amplitude after calibration of the weight sampling values at each time point of each sampling point.

[0016] Further, in Step 2, obtain the gradient between adjacent values in the sampling value queue of each sampling point, form a queue of the gradients between all adjacent values in the order of the sampling value queue, define it as the weight gradient queue of each sampling point, compare the values in the weight gradient queue of each sampling point with zero, and define the values with gradients higher than zero as , define the values with gradients lower than zero as , define the values with gradients equal to zero as , and thus obtain a queue formed by values of , define it as the weight gradient zero ratio queue of each sampling point.

[0017] Further, in Step 2, the calculation method of the trend coefficient of the weight sampling value at each time point in the sampling value queue of each sampling point is as follows:

[0018] Define the difference between each gradient in a partial interval at each time point in the weight gradient queue of each sampling point and the average of all gradients as the difference one of each gradient in the partial interval at each time point, and define the average of the difference one of all gradients in the partial interval at each time point as the partial difference two at each time point;

[0019] Obtain the gradient zero ratio coefficient at each time point according to the total count of the values corresponding in the weight gradient zero ratio queue in the partial interval at each time point;

[0020] Obtain the trend coefficient of the weight sampling value at each time point in the sampling value queue of each sampling point through the partial difference two and the gradient zero ratio coefficient at each time point;

[0021] Further, in Step 2, the calculation equation of the trend coefficient of the weight sampling value at each time point in the sampling value queue of each sampling point is as follows:

[0022]

[0023] In the equation, represents the th gradient in the partial interval at the th time point in the weight gradient queue of each sampling point, represents the average of all gradients in the partial interval at the th time point in the weight gradient queue of each sampling point, represents the value corresponding in the weight gradient zero ratio queue for the th gradient in the partial interval at the th time point, represents the number of all values in the partial interval at each time point, represents the trend coefficient of the weight sampling value at the th time point of each sampling point, is the Euler number.

[0024] Furthermore, in Step2, the operation method for the time-domain coefficient of the sampling value queue at each sampling point is as follows:

[0025] Obtain all the peak and valley points within the sampling value regression line. The peak and valley points include the peak value and the valley value. Combine the time intervals between all adjacent peak and valley points in the sampling value queue of each sampling point to form the time interval group of each sampling point. Define the difference between each value in the time interval group of each sampling point and the mean of all data as the difference three of each value in the time interval group of each sampling point. Define the value obtained by taking the reciprocal of the mean of the difference three of all values in the time interval group of each sampling point as the time-domain coefficient of the sampling value queue of each sampling point.

[0026] Furthermore, in Step2, the operation equation for the time-domain coefficient of the sampling value queue at each sampling point is:

[0027]

[0028] In the equation, represents the th time interval in the time interval group of each sampling point, represents the mean of all time intervals in the time interval group of each sampling point, represents the number of all time intervals in the time interval group of each sampling point, represents the time-domain coefficient of the sampling value queue of each sampling point, is the Euler number.

[0029] Furthermore, in Step2, the operation method for the sampling value perturbation amplitude at each time point of each sampling point is as follows:

[0030] Define the quantity obtained by multiplying the trend coefficient by the time-domain coefficient as the action coefficient one at each time point of each sampling point. Obtain the sampling value perturbation amplitude at each time point of each sampling point through the action coefficient one;

[0031] Furthermore, in Step2, the operation equation for the sampling value perturbation amplitude at each time point of each sampling point is:

[0032]

[0033] In the equation, represents the trend coefficient of the weighted sampling value at the th time point of each sampling point, represents the time-domain coefficient of the sampling value queue of each sampling point, represents the sampling value perturbation amplitude at the th time point of each sampling point.

[0034] Further, in Step 2, the operation method for the perturbation amplitude after calibration of the weight sampling values at each time point of each sampling point is as follows:

[0035] Define the mean of the similarities between the sampling value queues of each sampling point and all other sampling points as the arrangement difference attribute of the weight sampling values of each sampling point, perform an inverse processing on the arrangement difference attribute of the weight sampling values of each sampling point, and the value obtained from this inverse processing is the anomaly amplitude of each sampling point;

[0036] Calibrate the sampling value perturbation amplitude at each time point of each sampling point through the anomaly amplitude of each sampling point to obtain the perturbation amplitude after calibration of the weight sampling values at each time point of each sampling point.

[0037] Further, in Step 2, the operation equation for the perturbation amplitude after calibration of the weight sampling values at each time point of each sampling point is:

[0038]

[0039] In the equation, represents the sampling value perturbation amplitude at the th time point of the th sampling point, represents the perturbation amplitude after calibration of the weight sampling value at the th time point of the th sampling point, represents the anomaly amplitude of the th sampling point.

[0040] Further, in Step 2, the operation equation for the anomaly amplitude of each sampling point is:

[0041]

[0042] In the equation, represents the Pearson coefficient between the sampling value queues of the th sampling point and the th sampling point, represents the total number of all sampling points, represents the anomaly amplitude of the th sampling point, is the Euler number.

[0043] Further, in Step 3, obtain the weight sampling values of all the load cells at each time point, calculate the mean and variance of the weight sampling values of all the load cells at each time point, and through the mean and variance of the weight sampling values at each time point, obtain the normal distribution equation of the weight sampling values at each time point ; Obtain the weight sampling values of each sampling point at each time point in the normal distribution equation The corresponding equation value , and take as the typical coefficient of the weight sampling values of each sampling point at each time point.

[0044] Furthermore, in Step3, the calculation method of the weight of the weight sampling value of each sampling point at each time point is as follows:

[0045] Obtain the weight of the weight sampling value of each sampling point at each time point according to the typical coefficient of the weight sampling value of each sampling point at each time point and the calibrated disturbance amplitude.

[0046] Furthermore, in Step3, the calculation equation of the weight of the weight sampling value of each sampling point at each time point is:

[0047]

[0048] In the equation, represents the calibrated disturbance amplitude of the weight sampling value of the th sampling point at the th time point, represents the typical coefficient of the weight sampling value of the th sampling point at the th time point, represents the weight of the weight sampling value of the th sampling point at the th time point, is the Euler number.

[0049] Furthermore, in Step4, define the product obtained by multiplying the weight of the weight sampling value of each sampling point at each time point by the weight sampling value as the product value one at that time point, and take the average value obtained by adding the product value one at that time point of all sampling points and dividing by the number of sampling points as the calibrated weight value at that time point. The calibrated weight value at that time point is the weight value of the grinding machine at that time point.

[0050] A data processing device for intelligent grinding, comprising:

[0051] A position sensor, a reverse driver and a plurality of weighing sensors connected to the controller. The weighing sensors are used to sample the weight sampling value of the grinding machine and transmit the weight sampling value to the controller. The plurality of weighing sensors are respectively arranged at different sampling points at the driving end of the reverse driver. The position sensor is used to sample the position value of the grinding machine and transmit the position value to the controller. The controller is used to obtain the weight value of the grinding machine according to the weight sampling value transmitted by sampling, and then adjust the driving force output by the reverse driver according to the weight value and position value of the grinding machine;

[0052] The modules running on the controller include:

[0053] A sampling module, which is used to obtain the weight sampling values of several sampling points at the driving end of the reverse driver transmitted by sampling;

[0054] A disturbing module, which is used to obtain the trend coefficients of the weight sampling values at each time point in the sampling value queue of each sampling point through the change trend of the weight sampling values in the sampling value queue of each sampling point, and thus obtain the disturbing amplitude after calibration of the weight sampling values at each time point;

[0055] A weight value module, which is used to obtain the weight values of the weight sampling values at each time point of each sampling point through the arrangement of the weight sampling values of all sampling points at each time point at the driving end of the reverse driver and the disturbing amplitude after calibration;

[0056] A calibration module, which is used to calibrate the weight sampling values of all sampling points at each time point through the weight values of the weight sampling values at each time point of each sampling point, and obtain the weight value of the grinding machine.

[0057] The beneficial effects of the present invention are that, compared with the prior art, the technical effects of the present invention include:

[0058] According to the change trend of the weight sampling values in the weight inclination queue of each sampling point, obtain the trend coefficients of the weight sampling values at each time point in the weight value queue of each sampling point, and improve the accuracy of analyzing the disturbing amplitude of the weight value in terms of trend; through the difference in the time interval between all adjacent peak-valley points on the sampling value regression line of the weight value, obtain the time-domain coefficients of the weight sampling value queue of each sampling point, and improve the accuracy of analyzing the disturbing amplitude of the weight value in terms of time domain; through the trend coefficients and time-domain coefficients, obtain the disturbing amplitude of the weight sampling values at each time point of each sampling point; through the average of the similarities between the weight sampling value queues of each sampling point and other sampling points, calibrate the disturbing amplitude of the sampling value at each time point of each sampling point, and obtain the disturbing amplitude after calibration of the weight sampling values at each time point of each sampling point, improving the accuracy of analyzing the interference of the weight sampling values at each time point of each sampling point by external magnetic fields; through the arrangement of the weight sampling values of all sampling points at each time point, obtain the typical coefficients of the weight sampling values at each time point of each sampling point, and through the typical coefficients and the disturbing amplitude after calibration, obtain the weight values of the weight sampling values at each time point of each sampling point, improving the accuracy of the typicality of the weight sampling values at each time point of each sampling point; through the weight values of the weight sampling values at each time point of each sampling point, calibrate the weight sampling values of all sampling points at each time point, and obtain the weight value of the grinding machine, improving the accuracy of the weight value sampling of the grinding machine. Description of the Drawings

[0059] Figure 1 is a flowchart of the data processing method for intelligent grinding described in the present invention;

[0060] Figure 2 is a partial structural diagram of the data processing device for intelligent grinding described in the present invention. Detailed implementation manner

[0061] To make the objectives, technical solutions and advantages of the present invention clearer, the following will clearly and completely describe the technical solutions of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. The embodiments described herein are only part of the embodiments of the present invention, rather than all embodiments. According to the spirit of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0062] As Figure 1 shown, a data processing method for intelligent grinding described in the present invention includes:

[0063] The weighing sensor samples the weight sampling value of the grinding machine and transmits the weight sampling value to the controller, the position sensor samples the position value of the grinding machine and transmits the position value to the controller, and the controller obtains the weight value of the grinding machine according to the weight sampling value transmitted by sampling, and then adjusts the driving force output by the reverse driver according to the weight value and position value of the grinding machine;

[0064] The method by which the controller obtains the weight value of the grinding machine according to the weight sampling value transmitted by sampling includes:

[0065] Step1, obtain the weight sampling values of several sampling points at the driving end of the reverse driver transmitted by sampling;

[0066] To sample the weight sampling values of the sampling points on the driving end of the reverse driver, since there are differences between the weight sampling values of different sampling points, several weighing sensors are arranged at several sampling points on the driving end of the reverse driver, and the weight sampling values obtained by the weighing sensors at several sampling points are used to analyze and obtain the weight values of the sampling points of the grinding machine.

[0067] For example, ten weighing sensors are equidistantly arranged at the driving end of the reverse driver, the sampling frequency of the weighing sensor is one sampling per microsecond, and the weight sampling values of all time points on each weighing sensor in one second are obtained. The sampling time points of the weighing sensor and the position sensor are the same each time.

[0068] In a preferred but non-limiting embodiment of the present invention, in Step1, the weight sampling values of all time points at each sampling point are arranged in the order of their sampling time points to form a queue, which is defined as the sampling value queue of each sampling point.

[0069] Thus, a sampling value queue of all sampling points is obtained.

[0070] Step2: Based on the change trend of the weight sampling values in the sampling value queue of each sampling point, obtain the trend coefficient of the weight sampling values at each time point in the sampling value queue of each sampling point. Thus, until the disturbance amplitude after calibration of the weight sampling values at each time point is obtained;

[0071] In a preferred but non-limiting embodiment of the present invention, in Step2, the specific method for obtaining the disturbance amplitude after calibration of the weight sampling values at each time point includes: performing a regression process on the sampling value queue of each sampling point to obtain the sampling value regression line of each sampling point (that is, the sampling value regression line of the sampling value queue). By the difference in time intervals between all adjacent peak-valley points in the sampling value regression line (peak-valley points include peak values and valley values. A peak value is a coordinate point where the derivative on the sampling value regression line is zero, and the weight sampling value at this coordinate point on the sampling value regression line is greater than the weight sampling values of the two adjacent coordinate points on both sides of this coordinate point; a valley value is a coordinate point where the derivative on the sampling value regression line is zero, and the weight sampling value at this coordinate point on the sampling value regression line is smaller than the weight sampling values of the two adjacent coordinate points on both sides of this coordinate point), obtain the time-domain coefficient of the sampling value queue of each sampling point. Through the trend coefficient and the time-domain coefficient, obtain the sampling value disturbance amplitude at each time point of each sampling point. By the average of the similarities between the sampling value queues of each sampling point and other sampling points, calibrate the sampling value disturbance amplitude at each time point of each sampling point to obtain the disturbance amplitude after calibration of the weight sampling values at each time point of each sampling point.

[0072] Since there are some differences in the disturbance amplitudes of the external magnetic fields at different sampling points at the driving end of the reverse driver, and the components in the load cell at different sampling points are affected by different amounts of external magnetic fields, the disturbance amplitudes suffered by the weight sampling values at different sampling points are different. Therefore, the weight value of the grinding machine is calibrated based on the different disturbance amplitudes suffered by the weight sampling values at each sampling point.

[0073] When the polishing force of the polishing machine increases, the weight value of the polishing machine will gradually increase; as time goes by, the polishing force of the polishing machine will gradually decrease. Therefore, the weight sampling value has relative stability during the corresponding period. That is, during the polishing of the polishing machine, the polishing force often fluctuates regularly, causing the time-domain fluctuation of the weight value of the polishing machine. Such a fluctuation is relatively stable. If the weighing sensor is disturbed by an external magnetic field, the stability of its weight sampling value will be poor. Therefore, the disturbance amplitude of the external magnetic field can also be analyzed through the time-domain of the weight sampling value, and the weight sampling values of different sampling points caused by the corresponding disturbance amplitude at the driving end of the reverse driver also change in the same way with time.

[0074] In a preferred but non-limiting embodiment of the present invention, in Step 2, the gradient between adjacent values (the value is the weight sampling value) in the sampling value queue of each sampling point is obtained. The gradients between all adjacent values are formed into a queue in the order of the sampling value queue, and defined as the weight gradient queue of each sampling point. The value (the value is the gradient) of the weight gradient queue of each sampling point is compared with zero. The value with a gradient higher than zero is defined as , the value with a gradient lower than zero is defined as , the value with a gradient of zero is defined as , and thus a queue with a value of is formed and defined as the weight gradient zero ratio queue of each sampling point. Here, in the weight gradient queue, the weight gradient (the weight gradient is the gradient) at each time point (the time point is the sampling time point) is obtained based on the weight sampling value of the next adjacent time point of the weight sampling value at each time point. That is, the quotient obtained by dividing the amount obtained by subtracting the weight sampling value of this time point from the weight sampling value of the next adjacent time point by the amount obtained by subtracting this time point from the next adjacent time point. This quotient is the gradient at this time point, that is, the gradient between adjacent values in the sampling value queue of the sampling point.

[0075] When the disturbance amplitude of the external magnetic field on the weight sampling value of each sampling point is lower, the stability of the weight sampling value should be better. That is, correspondingly in the weight gradient queue, the difference between the gradients in each partial interval of each gradient is lower, and the characteristics of the gradient being higher than zero or not higher than zero are more consistent; when the disturbance amplitude of the external magnetic field on the weight sampling value of each sampling point is higher, the fluctuation of the weight sampling value of each sampling point is higher. That is, correspondingly in the weight gradient queue, the difference between the gradients in each partial interval of each gradient is higher, and the characteristics of the gradient being higher than zero or not higher than zero are more inconsistent. Therefore, through the change trend of the weight sampling value in the sampling value queue of each sampling point, the trend coefficient of the weight sampling value at each time point in the sampling value queue of each sampling point is obtained.

[0076] In addition, just as the values at each time point (these values can be gradients) are regarded as the center, the values adjacent to the center before this values and the values adjacent to the center after this values are regarded as the partial intervals of each center value, that is, the values in the partial intervals at each time point. Here, the center value is also the value in the partial interval. Here

[0077] In a preferred but non-limiting embodiment of the present invention, in Step2, the operation method of the trend coefficient of the weight sampling value at each time point in the sampling value queue of each sampling point is as follows:

[0078] Define the difference between the average of each gradient and the overall gradient in the partial interval at each time point in the weight gradient queue of each sampling point as the difference one of each gradient in the partial interval at each time point, and define the average of the difference one of the overall gradient in the partial interval at each time point as the partial difference two at each time point;

[0079] Obtain the gradient zero ratio coefficient at each time point according to the total amount of the values corresponding in the weight gradient zero ratio queue (this value is the element of the weight gradient zero ratio queue) in the partial interval at each time point;

[0080] Obtain the trend coefficient of the weight sampling value at each time point in the sampling value queue of each sampling point through the partial difference two and the gradient zero ratio coefficient at each time point;

[0081] Here, the partial difference two is inversely proportional to the trend coefficient of the weight sampling value, and the gradient zero ratio coefficient is directly proportional to the trend coefficient of the weight sampling value.

[0082] In a preferred but non-limiting embodiment of the present invention, in Step2, the operation equation of the trend coefficient of the weight sampling value at each time point in the sampling value queue of each sampling point is:

[0083]

[0084] In the equation, represents the th gradient in the partial interval at the th time point in the weight gradient queue of each sampling point, represents the average of the overall gradient in the partial interval at the th time point in the weight gradient queue of each sampling point, represents the th value corresponding to the th gradient in the partial interval at the th time point in the weight gradient zero ratio queue, The trend coefficient representing the weight sampling values at the th time point for each sampling point, is the Euler number. Here, adding one inside the equation is to avoid the sum of gradients being zero, which is not conducive to the quantity obtained by subsequent multiplication.

[0085] Here, represents the difference between the values in the weight gradient queue (this difference is the first difference). The lower this difference, the better the stability, that is, the higher the trend coefficient; otherwise, the stability is worse and the trend coefficient is lower. is part of the second difference; represents the total measurement of the values (these values are the elements of the weight gradient zero ratio queue) corresponding in the partial intervals at each time point in the weight gradient zero ratio queue. The higher the value, the better the stability, that is, the higher the trend coefficient; otherwise, the stability is worse and the trend coefficient is lower. is the gradient zero ratio coefficient at each time point.

[0086] Based on this, the trend coefficient of the weight sampling value at each time point in the sampling value queue of each sampling point is obtained.

[0087] When the weight sampling values at each sampling point are less disturbed by the external magnetic field, then the time-domain property (the time-domain property can be understood as orderliness) of the weight sampling values should be better; otherwise, the time-domain property of the weight sampling values should be worse. Here, the time-domain performance of the weight sampling values is analyzed by whether the time interval between the peak value and the valley value in the sampling value regression line of the sampling value queue is the same. Therefore, the time-domain coefficient of the sampling value queue of each sampling point is obtained through the difference in the time intervals between all adjacent peak-valley points in the sampling value regression line.

[0088] Just as the regression process is performed on the sampling value queue of each sampling point through the ridge regression method, that is, the ridge regression method is used to take the weight sampling values in the sampling value queue as the dependent variable (that is, the Y coordinate in the rectangular coordinate system), and the sampling time points of the weight sampling values in the sampling value queue as the independent variable (that is, the X coordinate in the rectangular coordinate system) to perform the regression process, so as to obtain the sampling value regression line of each sampling point.

[0089] In a preferred but non-limiting embodiment of the present invention, in Step2, the calculation method of the time-domain coefficient of the sampling value queue of each sampling point is as follows:

[0090] Obtain all the peak and valley points within the regression line of the sampled values. The peak and valley points include the peak values and the valley values. Combine the time intervals between all adjacent peak and valley points in the sampled value queue of each sampling point (this time interval is the value obtained by subtracting the sampling time point of the previous peak and valley point from the sampling time point of the latter peak and valley point among two adjacent peak and valley points) to form the time interval group of each sampling point. Define the difference between each value (this value is the time interval) in the time interval group of each sampling point and the mean of all data as the difference three of each value in the time interval group of each sampling point. Define the value obtained by inversely processing the mean of the differences three of all values in the time interval group of each sampling point as the time domain coefficient of the sampled value queue of each sampling point.

[0091] In a preferred but non-limiting embodiment of the present invention, in Step2, the operation equation of the time domain coefficient of the sampled value queue of each sampling point is:

[0092]

[0093] In the equation, represents the th time interval in the time interval group of each sampling point, represents the mean of all time intervals in the time interval group of each sampling point, represents the number of all time intervals in the time interval group of each sampling point, represents the time domain coefficient of the sampled value queue of each sampling point, is the Euler number.

[0094] Here, is the difference three, is the mean of the difference three, is the value obtained by inversely processing the mean of the differences three of all values in the time interval group of each sampling point. When the difference between the time intervals between all adjacent peak and valley points is lower, the time domain performance is better, that is, the time domain coefficient of the sampled value queue of each sampling point is higher; otherwise, the time domain performance is worse, that is, the time domain coefficient of the sampled value queue of each sampling point is lower.

[0095] And because when the weighted sampled value is disturbed by an external magnetic field, the stability and time domain performance of the sampled value queue will not be obvious, so through the trend coefficient and the time domain coefficient, the disturbance amplitude of the sampled value at each time point of each sampling point is obtained.

[0096] In a preferred but non-limiting embodiment of the present invention, in Step2, the operation method of the disturbance amplitude of the sampled value at each time point of each sampling point is:

[0097] Define the quantity obtained by multiplying the trend coefficient and the time domain coefficient as the action coefficient one of each time point of each sampling point, and obtain the disturbance amplitude of the sampled value at each time point of each sampling point through the action coefficient one;

[0098] Here, the effect coefficient 1 is inversely proportional to the disturbance amplitude of the sampling value at each time point of each sampling point.

[0099] In a preferred but non-limiting embodiment of the present invention, in Step 2, the calculation equation of the disturbance amplitude of the sampling value at each time point of each sampling point is:

[0100]

[0101] In the equation, Characterize each sampling point The trend coefficient of the weight sampling value at a time point, Characterize the time domain coefficient of the sample value queue of each sampling point, Characterize each sampling point The disturbance amplitude of the sampled value at a time point. Here, The introduction is to avoid is zero.

[0102] Here, when the trend coefficient and the time domain coefficient are higher, it means that the amplitude of disturbance caused by the external magnetic field is lower; when the trend coefficient and the time domain coefficient are lower, it means that the amplitude of disturbance caused by the external magnetic field is higher.

[0103] Based on this, the disturbance amplitude of the sampling value at each sampling point at each time point is obtained.

[0104] In addition, because the weight sampling values of different sampling points have the same trend change in the same period of time, that is, the weight sampling values of all sampling points increase or decrease together, the disturbance amplitude of the sampling values at each time point of each sampling point is calibrated by taking the average of the similarities between the sampling value queues of each sampling point and other sampling points, and the calibrated disturbance amplitude of the weight sampling values at each time point of each sampling point is obtained.

[0105] In a preferred but non-limiting embodiment of the present invention, in Step 2, the calculation method of the disturbance amplitude after the weight sampling value of each sampling point at each time point is:

[0106] The average of the similarities (which may be Peel's coefficient) between each sampling point and the sampling value queues of all other sampling points is defined as the arrangement distinguishing attribute of the weight sampling value of each sampling point, and an inverse treatment is performed on the arrangement distinguishing attribute of the weight sampling value of each sampling point. The value obtained by the inverse treatment is the abnormal amplitude of each sampling point;

[0107] The disturbance amplitude of the sampling value at each time point of each sampling point is corrected through the abnormal amplitude of each sampling point, and the corrected disturbance amplitude of the weight sampling value at each time point of each sampling point is obtained.

[0108] In a preferred but non-limiting embodiment of the present invention, in Step 2, the operation equation for the perturbation amplitude after calibration of the weight sampling values at each time point of each sampling point is:

[0109]

[0110] In the equation, represents the perturbation amplitude of the sampling value at the th time point of the th sampling point, represents the perturbation amplitude after calibration of the weight sampling value at the th time point of the th sampling point, represents the anomaly amplitude of the th sampling point.

[0111] Here, the higher the similarity, the lower the probability of anomaly generation, and the lower the Pearson coefficient, the higher the probability of anomaly generation.

[0112] In a preferred but non-limiting embodiment of the present invention, in Step 2, the operation equation for the anomaly amplitude of each sampling point is:

[0113]

[0114] In the equation, represents the Pearson coefficient between the sampling value queues of the th sampling point and the th sampling point, represents the total number of all sampling points, represents the anomaly amplitude of the th sampling point, is the Euler number.

[0115] Here, is the average similarity. The higher the Pearson coefficient, the lower the probability of anomaly generation, and the lower the Pearson coefficient, the higher the probability of anomaly generation.

[0116] Accordingly, the perturbation amplitude after calibration of the weight sampling values at each time point of each sampling point is obtained.

[0117] Step 3, through the arrangement and the perturbation amplitude after calibration of the weight sampling values of all sampling points at each time point on the driving end of the reverse driver, obtain the weights of the weight sampling values of all sampling points at each time point;

[0118] Because the arrangement of the weight sampling values at different sampling points on the driving end of the reverse driver is different, that is, the higher the weight sampling value obtained by the weighing sensor at some sampling points closer to the force application point, and the lower the weight sampling value obtained by the weighing sensor farther from the force application point. However, the weight sampling values of all sampling points should obtain the weight value of the grinding machine to analyze whether the weight sampling values of the sampling points on the driving end of the reverse driver exceed the abnormality. Therefore, by analyzing the weight sampling values of different sampling points, the typical amplitude of each sampling point is obtained, that is, the typicality of the weight sampling values of each sampling point is different. So, the weight value of the grinding machine is obtained through the weight sampling value and the typical amplitude of each sampling point, and then the driving force output by the reverse driver is adjusted by obtaining the new weight value and position value.

[0119] Because weight is formed by the spread of force, the higher the ratio of the same weight sampling values of all sampling points on the driving end of the reverse driver, the more it can reflect the typicality of the weight sampling values of the sampling points on the driving end of the reverse driver. Therefore, the typicality of the weight sampling values of each sampling point can be obtained through the arrangement or ratio of the weight sampling values of all sampling points.

[0120] In a preferred but non-limiting embodiment of the present invention, in Step3, the weight sampling values of all sampling points of weighing sensors at each time point are obtained, and the mean and variance of the weight sampling values of all sampling points at each time point are calculated. Through the mean and variance of the weight sampling values of all time points, the normal distribution equation of the weight sampling values at each time point is obtained ; the weight sampling values of each sampling point at each time point in the normal distribution equation are obtained and the corresponding equation values are obtained. Let be defined as the typical coefficient of the weight sampling value of each sampling point at each time point. The typical coefficient is the arrangement of the weight sampling values of all sampling points at each time point.

[0121] In a preferred but non-limiting embodiment of the present invention, in Step3, the calculation method of the weight of the weight sampling value of each sampling point at each time point is as follows:

[0122] According to the typical coefficient of the weight sampling value of each sampling point at each time point and the calibrated interference amplitude, the weight of the weight sampling value of each sampling point at each time point is obtained;

[0123] Here, the typical coefficient is directly proportional to the weight, and the calibrated interference amplitude is inversely proportional to the weight.

[0124] In a preferred but non-limiting embodiment of the present invention, in Step 3, the operation equation for the weights of the weight sampling values at each time point of each sampling point is as follows:

[0125]

[0126] In the equation, represents the disturbed amplitude after calibration of the weight sampling value at the -th time point of the -th sampling point, represents the typical coefficient of the weight sampling value at the -th time point of the -th sampling point, represents the weight of the weight sampling value at the -th time point of the -th sampling point, is the Euler number.

[0127] Here, when the typical coefficient of the weight sampling value at each time point of each sampling point is higher, the weight sampling value at that time point of that sampling point is more suitable for representing the weight value of the grinding machine, that is, the corresponding weight should be higher; when the disturbed amplitude of the sampled value after calibration at each time point of each sampling point is higher, the authenticity of the weight sampling value at that time point of that sampling point is lower, and the corresponding weight should be lower.

[0128] Accordingly, the weights of the weight sampling values at each time point of each sampling point are obtained.

[0129] Step 4: Through the weights of the weight sampling values at each time point of each sampling point, calibrate the weight sampling values of all sampling points at each time point to obtain the weight value of the grinding machine.

[0130] Since the weight sampling values of different sampling points on the driving end of the reverse driver are different, it is necessary to calibrate the weight sampling values of each sampling point at each time point through the weights of the weight sampling values at each time point of each sampling point to obtain the weight value of the grinding machine at each time point, so as to adjust the driving force output by the reverse driver according to the weight value and position value of the grinding machine.

[0131] In a preferred but non-limiting embodiment of the present invention, in Step 4, the product obtained by multiplying the weight of the weight sampling value at each time point of each sampling point by the weight sampling value is defined as the product value one at that time point, and the average value obtained by adding the product value one at that time point of all sampling points and dividing by the number of sampling points is used as the calibrated weight value at that time point, and the calibrated weight value at that time point is the weight value of the grinding machine at that time point.

[0132] As Figure 2 shown, a data processing device for intelligent grinding according to the present invention includes:

[0133] A position sensor, a reverse driver, and several weighing sensors connected to the controller. The weighing sensors are used to sample the weight sampling values of the grinding machine and transmit the weight sampling values to the controller. The several weighing sensors are respectively arranged at different sampling points preset at the driving end of the reverse driver. The position sensor is used to sample the position value of the grinding machine and transmit the position value to the controller. The controller is used to obtain the weight value of the grinding machine according to the weight sampling values transmitted by sampling, and then adjust the driving force output by the reverse driver according to the weight value and position value of the grinding machine;

[0134] The modules running on the controller include:

[0135] A sampling module, which is used to obtain the weight sampling values of several sampling points at the driving end of the reverse driver transmitted by sampling;

[0136] A disturbance module, which is used to obtain the trend coefficients of the weight sampling values at each time point in the sampling value queue of each sampling point through the change trend of the weight sampling values in the sampling value queue of each sampling point, and thus obtain the disturbance amplitude after calibration of the weight sampling values at each time point;

[0137] A weight value module, which is used to obtain the weight value of the weight sampling value at each time point of each sampling point through the arrangement of the weight sampling values of all sampling points at each time point at the driving end of the reverse driver and the disturbance amplitude after calibration;

[0138] A calibration module, which is used to calibrate the weight sampling values of all sampling points at each time point through the weight values of the weight sampling values at each time point of each sampling point, and obtain the weight value of the grinding machine.

[0139] The beneficial effects of the present invention are that, compared with the prior art, the technical effects of the present invention include:

[0140] According to the changing trend of the weight sampling values in the weight inclination queue of each sampling point, obtain the trend coefficients of the weight sampling values at each time point in the weight value queue of each sampling point, and improve the accuracy of analyzing the disturbance amplitude of the weight value in terms of trend; through the difference in the time intervals between all adjacent peak-valley points on the sampling value regression line of the weight value, obtain the time-domain coefficients of the weight sampling value queue of each sampling point, and improve the accuracy of analyzing the disturbance amplitude of the weight value in terms of time domain; through the trend coefficients and time-domain coefficients, obtain the disturbance amplitude of the weight sampling values at each time point of each sampling point; through the average of the similarities between the weight sampling value queues of each sampling point and other sampling points, calibrate the disturbance amplitude of the sampling values at each time point of each sampling point, and obtain the calibrated disturbance amplitude of the weight sampling values at each time point of each sampling point, improving the accuracy of analyzing the external magnetic field disturbance of the weight sampling values at each time point of each sampling point; through the arrangement of the weight sampling values of all sampling points at each time point, obtain the canonical coefficients of each sampling point at each time point, and through the canonical coefficients and the calibrated disturbance amplitude, obtain the weights of the weight sampling values at each time point of each sampling point, improving the accuracy of the typicality of the weight sampling values at each time point of each sampling point; through the weights of the weight sampling values at each time point of each sampling point, calibrate the weight sampling values of all sampling points at each time point, and obtain the weight value of the grinding machine, improving the accuracy of the weight value sampling of the grinding machine.

[0141] 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 them. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that: modifications or equivalent replacements can still be made to the specific implementation manners of the present invention, and any modifications or equivalent replacements that do 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 data processing method for intelligent grinding, characterized in that, Including: A weighing sensor samples the weight sampling value of the grinding machine and transmits the weight sampling value to the controller. A position sensor samples the position value of the grinding machine and transmits the position value to the controller. The controller obtains the weight value of the grinding machine based on the weight sampling value transmitted by sampling, and then adjusts the driving force output by the reverse driver according to the weight value and position value of the grinding machine. The method by which the controller obtains the weight value of the grinding machine based on the weight sampling value transmitted by sampling includes: Step1: Obtain the weight sampling values of several sampling points at the driving end of the reverse driver transmitted by sampling. Step2: Through the change trend of the weight sampling values in the sampling value queue of each sampling point, obtain the trend coefficient of the weight sampling values at each time point in the sampling value queue of each sampling point, and thus until obtaining the disturbance amplitude after calibration of the weight sampling values at each time point. Step3: Through the arrangement of the weight sampling values of all sampling points at each time point at the driving end of the reverse driver and the disturbance amplitude after calibration, obtain the weight of the weight sampling values at each time point of each sampling point. Step4: Through the weight of the weight sampling values at each time point of each sampling point, calibrate the weight sampling values of all sampling points at each time point to obtain the weight value of the grinding machine.

2. The data processing method for intelligent grinding according to claim 1, wherein, In Step1, arrange the weight sampling values of all time points at each sampling point in the order of their sampling time points to form a queue, which is defined as the sampling value queue of each sampling point.

3. The data processing method for intelligent grinding according to claim 2, wherein In Step2, the specific method for obtaining the disturbance amplitude after calibration of the weight sampling values at each time point includes: performing regression processing on the sampling value queue of each sampling point to obtain the sampling value regression line of each sampling point. Through the difference in the time interval between all adjacent peak-valley points in the sampling value regression line, obtain the time-domain coefficient of the sampling value queue of each sampling point. Through the trend coefficient and the time-domain coefficient, obtain the disturbance amplitude of the sampling values at each time point of each sampling point. Through the average of the similarities between the sampling value queues of each sampling point and other sampling points, calibrate the disturbance amplitude of the sampling values at each time point of each sampling point to obtain the disturbance amplitude after calibration of the weight sampling values at each time point of each sampling point.

4. The data processing method for intelligent grinding according to claim 3, wherein In Step 2, obtain the gradients between adjacent values in the sampling value queue of each sampling point, form a queue of all the gradients between adjacent values in the order of the sampling value queue, define it as the weight gradient queue of each sampling point, compare the values in the weight gradient queue of each sampling point with zero, and define the values with gradients higher than zero as , define the values with gradients lower than zero as , define the values with gradients equal to zero as , and thus obtain a queue formed by values of , and define it as the weight gradient zero ratio queue of each sampling point.

5. The data processing method for intelligent grinding according to claim 4, characterized in that, In Step2, the calculation method of the trend coefficient of the weight sampling values at each time point in the sampling value queue of each sampling point is: Define the difference between the inclination of each part interval at each time point in the weight inclination queue of each sampling point and the average of all inclinations as the difference one of the inclinations in each part interval at each time point. Define the average of the difference one of all inclinations in each part interval at each time point as the partial difference two at each time point. Obtain the inclination zero ratio coefficient at each time point according to the total measurement of the values corresponding to the weight inclination zero ratio queue in each part interval at each time point. Through the partial difference two and the inclination zero ratio coefficient at each time point, obtain the trend coefficient of the weight sampling values at each time point in the sampling value queue of each sampling point. In Step2, the calculation equation of the trend coefficient of the weight sampling values at each time point in the sampling value queue of each sampling point is: ； Within the equation, representing the th gradient in a partial interval at the th time point in the weight gradient queue for each sampling point, representing the mean of all gradients in a partial interval at the th time point in the weight gradient queue for each sampling point, representing the value corresponding to the th gradient in a partial interval at the th time point in the weight gradient zero ratio queue, representing the number of all values in a partial interval for each time point, representing the trend coefficient of the weight sampling value at the th time point for each sampling point, being the Euler number.

6. The data processing method for intelligent grinding according to claim 5, wherein In Step 2, the calculation method for the time-domain coefficient of the sampling value queue at each sampling point is as follows: Obtain all the peak and valley points within the sampling value regression line. The peak and valley points include the peak value and the valley value. Combine the time intervals between all adjacent peak and valley points in the sampling value queue of each sampling point to form the time interval group of each sampling point. Define the difference between each value in the time interval group of each sampling point and the mean of all data as the difference three of each value in the time interval group of each sampling point. Define the value obtained by taking the reciprocal of the mean of the difference three of all values in the time interval group of each sampling point as the time-domain coefficient of the sampling value queue of each sampling point; In Step 2, the calculation equation for the time-domain coefficient of the sampling value queue at each sampling point is: ； Within the equation, represents the th time interval within the time interval group of each sampling point, represents the mean of all time intervals in the time interval group of each sampling point, represents the number of all time intervals within the time interval group of each sampling point, represents the time-domain coefficient of the sampling value queue of each sampling point, is the Euler number; In Step 2, the calculation method for the sampling value perturbation amplitude at each time point of each sampling point is: Define the quantity obtained by multiplying the trend coefficient and the time-domain coefficient as the action coefficient one at each time point of each sampling point. Obtain the sampling value perturbation amplitude at each time point of each sampling point through the action coefficient one; In Step 2, the calculation equation for the sampling value perturbation amplitude at each time point of each sampling point is: ； Within the equation, The trend coefficient representing the weight sampling values at the th time point for each sampling point, The time-domain coefficient representing the sampling value queue of each sampling point, The sampling value disturbance amplitude at the th time point for each sampling point.

7. The data processing method for intelligent grinding according to claim 6, characterized in that, In Step 2, the calculation method for the perturbation amplitude of the weighted sampling value after calibration at each time point of each sampling point is: Define the mean of the similarities between the sampling value queue of each sampling point and the sampling value queues of all other sampling points as the arrangement difference attribute of the weighted sampling value of each sampling point. Perform an inverse processing on the arrangement difference attribute of the weighted sampling value of each sampling point. The value obtained from this inverse processing is the abnormal amplitude of each sampling point; Calibrate the sampling value perturbation amplitude at each time point of each sampling point through the abnormal amplitude of each sampling point to obtain the perturbation amplitude of the weighted sampling value after calibration at each time point of each sampling point; In Step 2, the calculation equation for the perturbation amplitude of the weighted sampling value after calibration at each time point of each sampling point is: ； Inside the equation, Characterize the th sampling point at the th time point of the sampling value disturbance amplitude, Characterize the th sampling point at the th time point of the disturbance amplitude after calibration of the weight sampling value, Characterize the th sampling point of the anomaly amplitude; In Step 2, the calculation equation for the abnormal amplitude of each sampling point is: ； Within the equation, represents the Pearson coefficient between the sampling value queues of the th sampling point and the represents the total number of sampling points, represents the anomaly amplitude of the th sampling point, is the Euler number.

8. The data processing method for intelligent grinding according to claim 7, wherein In Step 3, obtain the weight sampling values of all the sampling points at each time point, calculate the mean and variance of the weight sampling values of all the sampling points at each time point, and through the mean of the weight sampling values of all at each time point and the variance , obtain the normal distribution equation of the weight sampling values at each time point ; obtain the corresponding equation values of the weight sampling values of each sampling point at each time point within the normal distribution equation ; define as the canonical coefficient of the weight sampling values of each sampling point at each time point ; In Step 3, the calculation method for the weight of the weighted sampling value at each time point of each sampling point is: Obtain the weight of the weighted sampling value at each time point of each sampling point based on the typical coefficient and the calibrated perturbation amplitude of the weighted sampling value at each time point of each sampling point; In Step 3, the calculation equation for the weight of the weighted sampling value at each time point of each sampling point is: ； Inside the equation, representing the th weight sampling value of the th time point after calibration of the disturbance amplitude of the representing the th canonical coefficient of the weight sampling value of the th time point of the th sampling point, representing the th weight of the weight sampling value of the th time point of the is the Euler number.

9. The data processing method for intelligent grinding according to claim 8, wherein In Step 4, define the product obtained by multiplying the weight of the weighted sampling value at each time point of each sampling point by the weighted sampling value as the product value one at that time point. Take the mean of the sum of the product value one at that time point of all sampling points divided by the number of sampling points as the calibrated weight value at that time point. The calibrated weight value at that time point is the weight value of the grinding machine at that time point.

10. A data processing device for intelligent grinding, characterized in that, Include: A position sensor, a reverse driver, and a plurality of weighing sensors connected to the controller. The weighing sensors are used to sample the weight sampling value of the grinding machine and transmit the weight sampling value to the controller. The plurality of weighing sensors are respectively arranged at different sampling points at the driving end of the reverse driver. The position sensor is used to sample the position value of the grinding machine and transmit the position value to the controller. The controller is used to obtain the weight value of the grinding machine according to the weight sampling value transmitted by sampling, and then adjust the driving force output by the reverse driver according to the weight value and position value of the grinding machine; The modules running on the controller include: A sampling module, which is used to obtain the weight sampling values of a plurality of sampling points at the driving end of the reverse driver transmitted by sampling; A disturbance module, which is used to obtain the trend coefficient of the weight sampling values at each time point in the sampling value queue of each sampling point according to the change trend of the weight sampling values in the sampling value queue of each sampling point, and thus obtain the disturbance amplitude after calibration of the weight sampling values at each time point; A weight module, which is used to obtain the weight of the weight sampling values at each time point of each sampling point through the arrangement of the weight sampling values of all sampling points at each time point at the driving end of the reverse driver and the disturbance amplitude after calibration; A calibration module, which is used to calibrate the weight sampling values of all sampling points at each time point through the weights of the weight sampling values at each time point of each sampling point to obtain the weight value of the grinding machine.

Images