Numerical control machine tool control method and system based on multi-sensor fusion

Abstract

The invention relates to the field of equipment control, and provides a numerical control machine tool control method and system based on multi-sensor fusion. The method comprises the following steps: acquiring operation data of a numerical control machine tool through a plurality of controllers to obtain an initial sensor data set; inputting the initial sensor data set into a Kalman filter for noise filtering processing to obtain pure sensor data; performing multi-source data fusion processing through a weighted fusion algorithm according to the pure sensor data to obtain a fusion state vector; performing comparative analysis on the fusion state vector and a preset threshold matrix to obtain a machine tool operation state evaluation result; and generating a control instruction through a PID (Proportion Integration Differentiation) control algorithm according to the machine tool running state evaluation result, and sending the control instruction to an execution mechanism of the numerical control machine tool. By dynamically adjusting the fusion weight of the sensor and comprehensively judging multiple features, the influence of abnormal data is reduced, and the control stability is enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of equipment control, and in particular to a numerical control machine tool control method and system based on multi-sensor fusion. BACKGROUND

[0002] As the core equipment of modern manufacturing industry, the control precision and running stability of numerical control machine tools directly affect the machining quality of products. Traditional numerical control machine tool control systems mainly rely on single type sensors for state monitoring, such as monitoring tool position only through position encoders or monitoring spindle temperature only through temperature sensors. With the continuous improvement of machining precision requirements, multi-sensor fusion technology has been gradually introduced into the field of numerical control machine tool control. Existing multi-sensor fusion methods usually adopt simple data superposition or average processing methods to combine and analyze data from multiple sensors. These technologies can to some extent realize the comprehensive utilization of multi-sensor data and provide multi-dimensional state information for machine tool control.

[0003] However, the prior art has the following deficiencies: First, most schemes do not fully consider the reliability differences of each sensor data when performing sensor data fusion, and adopt fixed weights for data fusion, which leads to the fact that the fusion result is easily affected by contaminated data when some sensors are disturbed or fail, and the contribution of each sensor data cannot be dynamically adjusted; second, existing filtering processing is mainly performed independently for a single sensor, and lacks processing of time synchronization of multi-sensor data, and the time deviation of different sensors due to differences in sampling frequency and response time will reduce the accuracy of the fusion data; third, traditional state evaluation methods mostly adopt simple threshold judgment, and lack comprehensive consideration of the coupling relationship between multi-dimensional features such as vibration, temperature and position, and it is difficult to fully reflect the real running state of the machine tool. These problems make the existing multi-sensor fusion control scheme difficult to meet the needs of high-end manufacturing in terms of control precision and stability when facing complex machining conditions and variable working environments. SUMMARY

[0004] The present application provides a numerical control machine tool control method and system based on multi-sensor fusion to solve the defects of the prior art.

[0005] The first aspect of the present application provides a numerical control machine tool control method based on multi-sensor fusion, comprising: S1, acquiring running data of a numerical control machine tool through multiple controllers to obtain an initial sensor data set; S2, inputting the initial sensor data set into a Kalman filter for noise filtering processing to obtain pure sensor data; S3, performing multi-source data fusion processing on the pure sensor data through a weighted fusion algorithm to obtain a fusion state vector; S4. Compare and analyze the fused state vector with the preset threshold matrix to obtain the machine tool operating status evaluation result; S5. Based on the machine tool operating status evaluation results, a control command is generated using a PID control algorithm, and the control command is sent to the actuator of the CNC machine tool.

[0006] According to the multi-sensor fusion-based CNC machine tool control method provided by the present invention, the controller in step S1 includes: a vibration sensor, a temperature sensor, and a position sensor.

[0007] According to the multi-sensor fusion-based CNC machine tool control method provided by the present invention, step S2 further includes: S21. Perform timestamp alignment processing on the initial sensor dataset to obtain synchronized sensor data; S22. Input the synchronous sensor data into the Kalman filters corresponding to multiple sensors respectively, and perform recursive calculations through two stages of state prediction and measurement update to obtain the filtered data of multiple sensors. S23. Calculate the covariance matrix of the filtered data from multiple sensors to obtain clean sensor data.

[0008] According to the multi-sensor fusion-based CNC machine tool control method provided by the present invention, step S3 further includes: S31. Calculate the signal-to-noise ratio of multiple sensors based on the clean sensor data to obtain the reliability coefficient of multiple sensors; S32. Convert the reliability coefficients of multiple sensors into multiple fusion weight coefficients through normalization processing; S33. Perform a weighted summation operation on the pure sensor data and the corresponding fusion weight coefficients to obtain the fusion state vector.

[0009] According to the multi-sensor fusion-based CNC machine tool control method provided by the present invention, the calculation formula for the sensor reliability coefficient in step S31 is as follows:

[0010] in, For sensor index values, For the first The reliability coefficient of each sensor, For the first The variance of individual sensor data This is the preset minimum variance threshold.

[0011] According to the multi-sensor fusion-based CNC machine tool control method provided by the present invention, step S4 further includes: S41, extract the vibration amplitude feature, temperature change rate feature and position deviation feature from the fusion state vector, and obtain a feature parameter set; S42, compare multiple feature parameters in the feature parameter set with corresponding threshold ranges in a preset threshold matrix respectively, and obtain state identifiers of multiple features; S43, according to the state identifiers of multiple features, comprehensive evaluation is carried out through fuzzy inference rules, and a machine tool running state evaluation result is obtained.

[0012] According to the numerical control machine tool control method based on multi-sensor fusion provided by the application, step S5 further comprises: S51, determining a control target parameter according to the machine tool running state evaluation result; S52, calculating a proportional term, an integral term and a differential term through a PID control algorithm, and obtaining multiple control components; S53, summing multiple control components and performing amplitude limiting processing to obtain a control instruction; S54, sending the control instruction to a servo driver and a spindle controller through a communication bus.

[0013] The second aspect of the application further provides a numerical control machine tool control system based on multi-sensor fusion, comprising: A data acquisition module is configured to acquire running data of a numerical control machine tool through multiple controllers to obtain an initial sensor data set; A filtering processing module is configured to input the initial sensor data set into a Kalman filter to filter out noise and obtain pure sensor data; A data fusion module is configured to perform multi-source data fusion processing on the pure sensor data through a weighted fusion algorithm to obtain a fusion state vector; A state evaluation module is configured to compare and analyze the fusion state vector with a preset threshold matrix to obtain a machine tool running state evaluation result; A control execution module is configured to generate a control instruction through a PID control algorithm according to the machine tool running state evaluation result, and send the control instruction to an actuator of the numerical control machine tool.

[0014] The third aspect of the application further provides a numerical control machine tool control device based on multi-sensor fusion, comprising: A memory and at least one processor, wherein the memory stores instructions; At least one processor calls the instructions in the memory, so that a numerical control machine tool control device based on multi-sensor fusion executes a numerical control machine tool control method based on multi-sensor fusion as described in any one of the above.

[0015] The fourth aspect of the present application provides a computer readable storage medium, which stores instructions, and the instructions are executed by a processor to implement the multi-sensor fusion based numerical control machine tool control method according to any one of the above.

[0016] The present application can effectively eliminate random noise and system error generated in the measurement process by simultaneously acquiring multi-dimensional sensor information such as vibration, temperature and position in the data acquisition stage, and using Kalman filter to timestamp align and noise filter each sensor data, so that the data basis relied on subsequent processing is more pure and reliable; on this basis, by calculating the signal-to-noise ratio of each sensor data and converting it into a dynamic fusion weight coefficient, real-time evaluation and adaptive adjustment of the reliability of different sensor data are realized, when a certain sensor is disturbed by external interference or measurement anomaly, the system can automatically reduce the weight proportion of the sensor data, avoiding significant deviation of the overall fusion result caused by abnormal data, thereby ensuring the stability and accuracy of the fusion state vector; by extracting the vibration amplitude feature, temperature change rate feature and position deviation feature in the fusion state vector, and combining fuzzy inference rules for comprehensive judgment, the running state of the machine tool under different working conditions can be more comprehensively reflected, and the one-sidedness of single feature judgment is overcome; the finally generated control instruction is based on the comprehensive analysis result of multi-dimensional state information, so that the control decision is more targeted and timely, which can timely respond to the change of the machine tool running state, reduces the control deviation caused by lag or inaccuracy of state evaluation, and enhances the control stability and anti-interference ability of the numerical control machine tool in complex machining environment as a whole, providing more reliable technical support for ensuring machining precision and safe operation of equipment. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0018] Figure 1 A multi-sensor fusion based numerical control machine tool control method flowchart is provided for the embodiments of the present application. Figure 2 A multi-sensor fusion based numerical control machine tool control system structure diagram is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0019] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in combination with the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments, and they should not be understood as a limitation to the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application. In the description of the present application, it should be understood that the terms used are only for the purpose of description, and should not be understood as indicating or implying relative importance.

[0020] As shown in Figure 1 The present application provides a multi-sensor fusion-based numerical control machine tool control method, comprising: S1, collecting running data of the numerical control machine tool through multiple controllers to obtain an initial sensor data set.

[0021] Among them, the controller in step S1 includes a vibration sensor, a temperature sensor and a position sensor.

[0022] Further, during the running process of the numerical control machine tool, the vibration sensor is installed on the spindle box and the workbench position, and real-time collection of three-axis vibration signals generated by the machine tool during cutting processing is performed, with a sampling frequency of 1000Hz, and each sampling point containing X, Y and Z three direction acceleration values; the temperature sensor is distributed on the spindle bearing, the feeding motor and the cooling system key node, and the temperature readings of each monitoring point are collected at a frequency of 100Hz; the position sensor obtains the position coordinate information of each motion axis in real time through the grating ruler and the rotary encoder, and the sampling frequency is 500Hz. The three types of sensors continuously collect data within the same time period, and finally all sensor data is collected to the data acquisition unit through the field bus to form an initial sensor data set, which contains all original measurement values of the three types of sensors within a continuous time window.

[0023] S2, inputting the initial sensor data set into a Kalman filter for noise filtering processing to obtain pure sensor data.

[0024] Among them, step S2 further comprises: S21, performing timestamp alignment processing on the initial sensor data set to obtain synchronized sensor data.

[0025] In step S21, in the initial sensor data set, the data timestamps of different sensors are inconsistent due to the difference in sampling frequency. Specifically, the present application selects the lowest sampling frequency of 100 Hz as the reference time interval of 10 ms, and resamples all sensor data according to the time axis. For 1000 Hz data of the vibration sensor, the average value of every 10 consecutive sampling points is calculated as the representative value of the 10 ms time window; for 500 Hz data of the position sensor, the average value of every 5 consecutive sampling points is calculated; and the 100 Hz data of the temperature sensor is directly retained. Subsequently, the present application establishes a unified time axis, the starting point of which is the start time of data collection, and each time node is spaced 10 ms apart, and then the data of each sensor after down-sampling or retention is arranged in sequence according to the time axis, ensuring that there are corresponding measurement values of the three types of sensors at each 10 ms time node, forming a time-synchronized data matrix.

[0026] S22, input the synchronized sensor data into the Kalman filter corresponding to each sensor respectively, and perform recursive operation through two stages of state prediction and measurement update to obtain the filtered data of the multiple sensors.

[0027] In step S22, for the vibration sensor data, the present application establishes state equations and observation equations, the state variables of which include vibration acceleration and vibration speed, and the system matrix is determined according to the vibration mechanics model. The Kalman filter first performs state prediction, calculates the state prediction value at the current time according to the state estimation value at the previous time and the state transition matrix, and updates the prediction error covariance matrix according to the process noise covariance; then performs measurement update to calculate the Kalman gain, which reflects the weight distribution relationship between the prediction value and the actual measurement value, and the gain value is determined by the prediction error covariance and the measurement noise covariance. Subsequently, the present application multiplies the difference between the actual measured vibration acceleration value and the prediction value by the Kalman gain to obtain the state correction amount, adds the correction amount to the predicted state to obtain the optimal state estimation value at the current time. Then, the temperature sensor and the position sensor data are input into independent Kalman filters respectively, and the same prediction-update recursive process is performed, the state variables of the temperature filter are temperature value and temperature change rate, and the state variables of the position filter are position coordinates and movement speed, the three filters run in parallel, and output their respective filtered data sequences.

[0028] S23, perform covariance matrix calculation on the filtered data of the multiple sensors to obtain pure sensor data.

[0029] Furthermore, in step S23, this invention calculates the covariance matrix of the filtered data sequence from the vibration sensor. The diagonal elements of this matrix represent the data variance of each measurement dimension, and the off-diagonal elements represent the correlation between different dimensions. Specifically, the vibration acceleration data at N consecutive time points after filtering is taken, with N set to 100. The mean values ​​of the accelerations in the X, Y, and Z directions are calculated, and then the sum of squared deviations of each data point from the mean is calculated and divided by the sample size to obtain the variance value. Subsequently, the covariance matrices of the temperature sensor and the position sensor are calculated using the same method, respectively, for their respective measurement dimensions. The magnitude of the covariance matrix reflects the degree of fluctuation in the filtered data; a smaller value indicates a better filtering effect. Finally, this invention combines the filtered data and their covariance matrices from the three types of sensors to form clean sensor data, where the filtered data consists of the measured values ​​at each time point, and the covariance matrix is ​​stored additionally as a quantitative indicator of data quality.

[0030] S3. Based on the pure sensor data, perform multi-source data fusion processing using a weighted fusion algorithm to obtain a fusion state vector.

[0031] Step S3 further includes: S31. Calculate the signal-to-noise ratio of multiple sensors based on the clean sensor data to obtain the reliability coefficient of multiple sensors.

[0032] The formula for calculating the sensor reliability coefficient in step S31 is as follows:

[0033] in, For sensor index values, For the first The reliability coefficient of each sensor, For the first The variance of individual sensor data This is the preset minimum variance threshold.

[0034] In step S31, this invention extracts the data variance of each sensor within a specific time window from the clean sensor data. This variance value is directly derived from the diagonal elements of the covariance matrix calculated in step S23. For the vibration sensor, the average of the variances in the X, Y, and Z directions is taken as the overall variance of the sensor. This invention sets a minimum variance threshold of 0.001, which is pre-calibrated based on the sensor's physical characteristics and measurement accuracy. When calculating the sensor reliability coefficient, the sensor data variance is divided by the sum of the variance and the threshold. The smaller the variance value, the larger the reliability coefficient of the calculated result, with the value approaching 1; the larger the variance value, the smaller the reliability coefficient, with the value approaching 0. Finally, the reliability coefficients of the three types of sensors are calculated separately, and the reliability coefficient of the vibration sensor is denoted as... The reliability coefficient of the temperature sensor is denoted as The reliability coefficient of the position sensor is denoted as .

[0035] S32. The reliability coefficients of multiple sensors are converted into multiple fusion weight coefficients through normalization processing.

[0036] This invention sums the reliability coefficients of the three sensors to obtain the total. Then, the reliability coefficients of the three sensors are divided by the sum. The fusion weighting coefficients for the vibration sensor, temperature sensor, and position sensor are obtained. After normalization, the sum of the three weighting coefficients equals 1, and the value of each weighting coefficient ranges from 0 to 1. When the variance of a sensor's data increases, its reliability coefficient decreases. After normalization, the weighting coefficient corresponding to that sensor decreases, while the weighting coefficients of other sensors increase accordingly, achieving dynamic adjustment of the weights.

[0037] S33. Perform a weighted summation operation on the pure sensor data and the corresponding fusion weight coefficients to obtain the fusion state vector.

[0038] This invention multiplies the filtered data from the vibration sensor, temperature sensor, and position sensor by their respective weighting coefficients, and then sums the three products element-wise. Since the three types of sensors measure different physical quantities, this invention normalizes the dimensions of each sensor's data before weighted summation. Vibration acceleration data is divided by a preset maximum acceleration value, temperature data is subtracted from a reference temperature and then divided by the temperature range, and position data is divided by the maximum travel distance, ensuring that the numerical range of each data type is uniformly between 0 and 1. The normalized vibration, temperature, and position data are then multiplied by their corresponding weighting coefficients and summed to obtain the individual element value of the fused state vector. This calculation process is repeated at each time point, forming a time-varying sequence of fused state vectors. Each element of the vector comprehensively reflects the combined measurement results of the three types of sensors at that moment.

[0039] S4. Compare and analyze the fused state vector with the preset threshold matrix to obtain the machine tool operating status evaluation result.

[0040] Step S4 further includes: S41. Extract vibration amplitude features, temperature change rate features, and position deviation features from the fused state vector to obtain a set of feature parameters.

[0041] Further, in step S41, in the vibration amplitude feature extraction, first, the vibration component of the fusion state vector in the time window is subjected to fast Fourier transform, the time domain signal is converted into the frequency domain spectrum, the frequency component with the largest amplitude in the spectrum is identified, and the amplitude is taken as the vibration amplitude feature value; in the temperature change rate feature extraction, the temperature components of the fusion state vector at the start point and the end point of the time window are taken, and the difference value divided by 1 second is taken as the temperature change rate feature value; in the position deviation feature extraction, the command positions of each axis are obtained from the numerical control system, the position component of the fusion state vector is subtracted from the command position axis by axis, and the maximum value of the absolute values of the X, Y and Z three-axis deviations is taken as the position deviation feature value. Finally, the three feature values form a feature parameter set.

[0042] S42, compare the plurality of feature parameters in the feature parameter set with the corresponding threshold value range in the preset threshold value matrix respectively, and obtain the state identification of the plurality of features.

[0043] Further, the preset threshold value matrix contains three rows and three columns, each row corresponds to the normal, warning and abnormal three interval thresholds of one feature parameter. In a specific embodiment, the vibration amplitude threshold is set to 0 to 0.3, 0.3 to 0.6 and 0.6 or more; the temperature change rate threshold is set to -0.1 to 0.1, -0.2 to -0.1, 0.1 to 0.2, less than -0.2 or greater than 0.2; the position deviation threshold is set to 0 to 0.05, 0.05 to 0.1 and 0.1 or more. The vibration amplitude value is compared with the threshold interval of the first row, and falls into the normal range, marked as "0", the warning range marked as "1", and the abnormal range marked as "2". The temperature change rate and the position deviation are compared in the same way to generate the state identification. The three state identifications form a state identification vector.

[0044] S43, according to the state identification of the plurality of features, comprehensive evaluation is carried out through fuzzy inference rules to obtain the machine tool running state evaluation result.

[0045] The present application defines 27 fuzzy inference rules in advance, and the rule format is "if the vibration state identification is X and the temperature state identification is Y and the position state identification is Z, then the machine tool running state is W". The state identification vector is matched with the 27 rules one by one, and the completely consistent rule is found. For example, the state identification vector {0, 0, 0} matches the rule output "normal", the state identification vector {1, 0, 0} matches the rule output "mild abnormality", and the state identification vector {2, 2, 1} matches the rule output "serious abnormality". Finally, the present application outputs the machine tool running state evaluation result according to the matching rule, and the result is one of the four levels of normal, mild abnormality, moderate abnormality and serious abnormality.

[0046] S5, generating a control instruction according to the machine tool running state evaluation result through a PID control algorithm, and sending the control instruction to the actuator of the numerical control machine tool.

[0047] wherein, the step S5 further comprises: S51, determining the control target parameters according to the machine tool operation state evaluation result.

[0048] In step S51, the application determines the control strategy according to the evaluation result. When the evaluation result is normal, the target position deviation is set to 0, and the current machining parameters are maintained; when the evaluation result is mild abnormality, the target position deviation is set to 0, the feed speed is reduced to 90% of the rated value, and the target vibration amplitude is set to 80% of the current value; when the evaluation result is moderate abnormality, the target position deviation is set to 0, the feed speed is reduced to 70% of the rated value, and the spindle speed is reduced to 85% of the rated value; when the evaluation result is severe abnormality, the feed motion is paused, and the spindle speed is reduced to 50% of the rated value. Finally, the application forms a control target parameter set consisting of the target position deviation, the target feed speed, and the target spindle speed.

[0049] S52, calculating the proportional term, the integral term, and the differential term through the PID control algorithm to obtain a plurality of control components.

[0050] In step S52, the application first extracts the actual position value from the fusion state vector, subtracts the command position to obtain the position error, and multiplies the error by the proportional coefficient to obtain the proportional term. An error accumulation variable is established, and the current error is accumulated every control period. The accumulation variable is multiplied by the integral coefficient and the control period to obtain the integral term. Subsequently, the error of the last period is recorded, and the error of the last period is subtracted from the current error to obtain the error change. The change is divided by Δt and multiplied by the differential coefficient to obtain the differential term. Finally, the feed speed control and the spindle speed control are executed independently PID calculation to obtain their respective three control components.

[0051] S53, summing a plurality of control components and performing amplitude limiting processing to obtain a control command.

[0052] Further, in step S53, the application performs arithmetic summing of the proportional term, the integral term, and the differential term of the position control to obtain the position control output , and sets the upper limit of the amplitude to 100 and the lower limit to -100. When is greater than 100, it is forcibly set to 100, and when it is less than -100, it is forcibly set to -100, and the original value is maintained in the interval. The feed speed control and the spindle speed control are summed and amplitude limited in the same way to obtain and . Finally, the application encapsulates the three control commands into a data packet in the format of {command type, target axis number, control amount value}, and the position control command contains independent control amounts of X, Y, and Z axes to form a complete control command.

[0053] S54, send the control instruction to the servo driver and spindle controller through the communication bus.

[0054] In step S54, the application sends the control instruction through the EtherCAT industrial Ethernet bus, the position control instruction is sent to the servo driver, X-axis address 01, Y-axis address 02, Z-axis address 03. Specifically, the application constructs an EtherCAT data frame, the frame header contains the target device address, the data field contains the instruction type, the axis number, and the control value, and the frame tail contains the check code. The servo driver receives and analyzes the control value as the current loop or speed loop input to drive the motor. The feed speed control instruction is sent to the feed shaft servo driver, and the spindle speed control instruction is sent to the address 10 spindle controller. The control period is 10ms, and three types of control instructions are sent in each period, and the actuator executes the instruction to adjust the motor output.

[0055] As shown in Figure 2 The application also provides a numerical control machine tool control system based on multi-sensor fusion, comprising: A data acquisition module 100 is configured to acquire running data of the numerical control machine tool through a plurality of controllers to obtain an initial sensor data set; A filtering processing module 200 is configured to input the initial sensor data set into a Kalman filter for noise filtering processing to obtain pure sensor data; A data fusion module 300 is configured to perform multi-source data fusion processing on the pure sensor data through a weighted fusion algorithm to obtain a fusion state vector; A state evaluation module 400 is configured to compare and analyze the fusion state vector with a preset threshold matrix to obtain a machine tool running state evaluation result; A control execution module 500 is configured to generate a control instruction through a PID control algorithm according to the machine tool running state evaluation result and send the control instruction to an actuator of the numerical control machine tool.

[0056] The application also provides a numerical control machine tool control device based on multi-sensor fusion, comprising a memory and at least one processor, wherein the memory stores instructions; and the at least one processor invokes the instructions in the memory to enable the numerical control machine tool control device based on multi-sensor fusion to perform the numerical control machine tool control method based on multi-sensor fusion according to any one of the above.

[0057] The application also provides a computer readable storage medium, wherein the computer readable storage medium stores instructions, and the instructions are executed by a processor to implement the numerical control machine tool control method based on multi-sensor fusion according to any one of the above.

[0058] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e., can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0059] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be realized by means of software and the necessary general hardware platform, and of course can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, which can be stored in a computer readable storage medium such as ROM / RAM, magnetic disk, optical disk, etc., and includes a plurality of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute a multi-sensor fusion-based numerical control machine tool control method.

[0060] The present application eliminates the influence of measurement noise and time deviation on data quality by time stamp alignment and Kalman filtering of multi-sensor data, and obtains reliable pure sensor data; real-time evaluation of sensor reliability is realized by calculating the variance of each sensor data and converting it into a dynamic fusion weight coefficient, when a sensor is disturbed, its weight proportion is automatically reduced to avoid abnormal data pollution of the fusion result; the limitations of single feature judgment are overcome by extracting three types of features of vibration amplitude, temperature change rate and position deviation and combining fuzzy inference rules for comprehensive judgment, and the real running state of the machine tool is fully reflected; according to the evaluation result, the PID algorithm is used to generate targeted control instructions, realizing collaborative control based on multi-dimensional state information, reducing control deviation, and enhancing the stability and anti-interference ability of the numerical control machine tool under complex working conditions.

[0061] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A multi-sensor fusion based control method for a CNC machine tool, characterized by, The method comprises the following steps: S1, collecting operation data of the numerical control machine tool through multiple controllers to obtain an initial sensor data set; S2, inputting the initial sensor data set into a Kalman filter for noise filtering processing to obtain pure sensor data; S3, performing multi-source data fusion processing on the pure sensor data through a weighted fusion algorithm to obtain a fusion state vector; S4, comparing and analyzing the fusion state vector with a preset threshold matrix to obtain a machine tool operation state evaluation result; S5, generating a control instruction through a PID control algorithm according to the machine tool operation state evaluation result, and sending the control instruction to an executing mechanism of the numerical control machine tool.

2. The multi-sensor fusion based control method of a CNC machine tool according to claim 1, characterized in that, The controller in step S1 comprises a vibration sensor, a temperature sensor and a position sensor.

3. The multi-sensor fusion based control method of a CNC machine tool according to claim 1, wherein, Step S2 further comprises: S21, timestamp alignment processing is performed on the initial sensor data set to obtain synchronized sensor data; S22, the synchronized sensor data is input into multiple sensor corresponding Kalman filters respectively, and recursive operation is performed through state prediction and measurement update two stages to obtain filter data of multiple sensors; S23, covariance matrix calculation is performed on the filter data of multiple sensors to obtain pure sensor data.

4. The multi-sensor fusion based control method for CNC machine tools according to claim 1, wherein, Step S3 further comprises: S31, the signal-to-noise ratio of multiple sensors is calculated according to the pure sensor data to obtain multiple sensor reliability coefficients; S32, multiple fusion weight coefficients are converted from multiple sensor reliability coefficients through normalization processing; S33, weighted summation operation is performed on the pure sensor data and the corresponding fusion weight coefficients to obtain a fusion state vector.

5. The multi-sensor fusion based control method of a CNC machine tool according to claim 3, wherein, The calculation formula of the sensor reliability coefficient in step S31 is: ; wherein, is a sensor index value, is a reliability coefficient of the sensor, is a variance of the sensor data, is a preset minimum variance threshold value.

6. The multi-sensor fusion based control method of a CNC machine tool according to claim 1, wherein, Step S4 further comprises: S41, vibration amplitude features, temperature change rate features and position deviation features are extracted from the fusion state vector to obtain a feature parameter set; S42, multiple feature parameters in the feature parameter set are compared with corresponding threshold ranges in the preset threshold matrix respectively to obtain state identifiers of multiple features; S43, according to the state identifiers of multiple features, comprehensive judgment is performed through fuzzy inference rules to obtain a machine tool operation state evaluation result.

7. The multi-sensor fusion based control method of a CNC machine tool according to claim 1, wherein, Step S5 further comprises: S51, determining a control target parameter according to the machine tool operation state evaluation result; S52, calculating a proportional term, an integral term and a differential term through a PID control algorithm to obtain multiple control components; S53, performing summation operation on multiple control components and performing amplitude limiting processing to obtain a control instruction; S54, sending the control instruction to a servo driver and a spindle controller through a communication bus.

8. A multi-sensor fusion based control system for a CNC machine tool, characterized by, The method comprises the following steps: A data acquisition module is configured to collect operation data of the numerical control machine tool through multiple controllers to obtain an initial sensor data set; A filtering processing module is configured to input the initial sensor data set into a Kalman filter for noise filtering processing to obtain pure sensor data; A data fusion module is configured to perform multi-source data fusion processing on the pure sensor data through a weighted fusion algorithm to obtain a fusion state vector; A state evaluation module is configured to compare and analyze the fusion state vector with a preset threshold matrix to obtain a machine tool operation state evaluation result. A control execution module is configured to generate a control instruction through a PID control algorithm according to the machine tool operation state evaluation result, and send the control instruction to an execution mechanism of the numerical control machine tool.

9. A multi-sensor fusion based control device for a CNC machine tool, characterized by, Comprise: a memory and at least one processor, wherein the memory stores instructions; at least one processor calls the instructions in the memory, so that a numerical control machine tool control device based on multi-sensor fusion executes a numerical control machine tool control method based on multi-sensor fusion as claimed in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores instructions, and the instructions are executed by the processor to realize a numerical control machine tool control method based on multi-sensor fusion as claimed in any one of claims 1-7.

Images