CN110568761A - On-line Optimizing Method of Feed Speed Based on Fuzzy Control - Google Patents

Abstract

The invention discloses an online optimization method of feed speed based on fuzzy control, which is used to solve the technical problem of poor practicability of the existing online feed speed control method. The technical solution is to realize the constant power control of the cutting process for complex and changeable control systems that are difficult to express with precise mathematical models by making a constant power fuzzy controller and applying fuzzy control algorithms. On-line optimization of the process parameters, to achieve constant power adaptive on-line control, to avoid excessive wear of the tool and vibration of the machine tool spindle under variable working conditions, to shorten the processing time and improve the processing efficiency under the premise of protecting the machine tool and the tool The role, and has wide applicability and practical engineering application value, good practicability.

Description

On-line Optimizing Method of Feed Speed Based on Fuzzy Control

technical field

The invention relates to an online feed speed control method, in particular to an online feed speed optimization method based on fuzzy control.

Background technique

In the current era of rapid development of CNC machining technology, the optimization of machining process parameters is helpful to give full play to the optimal driving performance of machine tools and the best cutting performance of cutting tools, thereby improving processing quality, reducing costs, and improving energy efficiency. There are not many process parameters that can be optimized in actual NC machining, such as adjusting the spindle speed or feed rate to optimize the machining process. However, the spindle speed reflects the movement of the machine tool spindle, and frequent changes will cause damage to the machine tool. Therefore, the optimization of the feed rate is the most direct and effective way to optimize the process parameters and improve the processing efficiency.

The literature "On-line control of side milling deformation of thin-walled parts based on finite element numerical model and feed rate optimization, Chinese Journal of Mechanical Engineering, 2017, Vol.53, No.21, p190-199" proposes a method based on finite element numerical model and advanced Based on the online control method of speed optimization, according to the finite element simulation results of the cutting process of thin-walled parts, the numerical model among the feed speed, cutting force and workpiece cutting deformation of CNC machine tools is established, and then the optimal target cutting force for controlling deformation is determined . On the open and modularized CNC system platform, the real-time acquisition of cutting force signal, filtering function and online optimization strategy of feed speed are developed, and the feed speed of the machine tool is adjusted in real time according to the filtered cutting force and corresponding algorithm during the machining process. The on-line control of cutting deformation is realized by ensuring that the cutting force is gradually approaching the optimal control target. The numerical model between feed speed and cutting force established in this document is obtained through finite element simulation, so its optimization effect depends on the accuracy of numerical simulation, and the numerical model does not consider processing parameters such as cutting width and cutting depth. It is suitable for multi-axis CNC machining of complex surfaces, and its application is subject to restrictions.

Contents of the invention

In order to overcome the disadvantage of poor practicability of the existing online feed speed control method, the present invention provides an online feed speed optimization method based on fuzzy control. By making a constant power fuzzy controller and applying a fuzzy control algorithm, the method realizes constant power control of the cutting process for a complex and changeable control system that is difficult to express with an accurate mathematical model. On-line optimization of process parameters, to achieve constant power self-adaptive on-line control, to avoid excessive tool wear and vibration of the machine tool spindle under variable working conditions, to shorten the processing time and improve the processing efficiency under the premise of protecting the machine tool and tools Function, and has wide applicability and practical engineering application value, good practicability.

The technical solution adopted by the present invention to solve its technical problems: a method for online optimization of feed speed based on fuzzy control, which is characterized in that it includes the following steps:

Step 1: Determine the input and output of the fuzzy controller. Set the target cutting power P _obj , collect the spindle power of the machine tool during processing, compare the actual processing power with the target power value P _obj to obtain the power error E _P , and at the same time differentiate the error, and calculate the power error in one sampling cycle Change rate C _P , power error E _P and power error change rate C _P are used as the input variables of the fuzzy controller, and the feed override of the CNC machine tool is used as the output.

Step 2, input and output fuzzy processing. Carry out corresponding fuzzy processing on the input and output precise value variables, using seven words to describe, namely {negative large, negative medium, negative small, zero, positive small, positive medium, positive large}, the input and output variables of fuzzy control are set as follows:

The fuzzy set of Ep is {NB, NM, NS, N0, 0, PS, PM, PB};

The fuzzy set of Cp is {NB, NM, NS, N0, 0, PS, PM, PB};

The fuzzy set of △U is {NB, NM, NS, 0, PS, PM, PB};

The domains of Ep, Cp, and △U are all {-6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6};

Step 3: Determine the membership function of the fuzzy subset. The input variable uses a triangular membership function, and the output variable uses a trapezoidal membership function.

Step four, formulate fuzzy control rules. Use If...then conditional statements to control the rules.

Step 5. Using the defuzzification method of the center of gravity method, the output fuzzy set is converted into a definite numerical output, and a fuzzy lookup table is made, and the corresponding output control quantity is obtained through the lookup table for a given input.

Step six, determine the parameters of the fuzzy controller. Determine the change interval of the input spindle power, the change interval of the spindle power change rate, and the change interval of the output feedrate override. At the same time, it is necessary to choose a reasonable scale factor and quantization factor. When the domain of error is set to [-x,+x], the fuzzy set domain of error is {-n, -n+1,...,0,...,n -1, n}, the quantization factor K of the error is expressed as follows:

Among them, n is the maximum value of the discourse domain of the error fuzzy set, and x is the maximum value of the discourse domain of the error.

Step seven, making fuzzy control lookup table. Make a simple fuzzy controller in the SIMULINK module of MATLAB software, and load the previously established fuzzy rules into the workspace, map the input and output variables to the corresponding input and output of the fuzzy controller in the system test, run the test and save the results, Transform the result to get the corresponding fuzzy control query table.

Step eight, replace the fuzzy controller in the online optimization process with the fuzzy control look-up table, and establish an online optimization model based on fuzzy control, which is built into the CNC machine tool through programming.

Step nine, online optimization and debugging. Collect real-time milling power data through CNC milling workpieces, use the online optimization controller to adaptively adjust the feed rate and then adjust the feed speed, and feed the adjusted power back to the constant power controller for continuous iterative adjustment until the power is constant .

The beneficial effects of the present invention are: the method realizes the constant power control of the cutting process for complex and changeable control systems that are difficult to express with precise mathematical models by making a constant power fuzzy controller and applying the fuzzy control algorithm. The present invention can solve complex curved surfaces On-line optimization of process parameters in the multi-axis CNC milling process realizes constant power self-adaptive on-line control, which ensures the avoidance of excessive tool wear and vibration of the machine tool spindle under variable working conditions, and shortens the time while protecting the machine tool and the tool. It can reduce the processing time and improve the processing efficiency, and has wide applicability and practical engineering application value, and has good practicability.

The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

Description of drawings

Fig. 1 is a flow chart of the online optimization method of feed speed based on fuzzy control in the present invention.

Fig. 2 is a graph of membership function of input and output variables in the method of the present invention.

Fig. 3 is a simulation model diagram of online optimization of feed speed based on fuzzy control in the method of the present invention.

Fig. 4 is a comparison chart of online optimization simulation results of feed speed based on fuzzy control in the method of the present invention.

Detailed ways

Refer to Figure 1-4. The present invention is based on fuzzy control feed speed online optimization method concrete steps as follows:

Step 1: Determine the input and output of the fuzzy controller. Set the target cutting power P _obj , collect the spindle power of the machine tool during processing, compare the actual processing power with the target power value P _obj to obtain the power error E _P , and differentiate the error to calculate the power in one sampling period The error change rate C _P , therefore, the power error E _P and the power error change rate C _P are used as the input variables of the fuzzy controller, and the feed override of the CNC machine tool is taken as the output.

Step 2, input and output fuzzy processing. Corresponding fuzzy processing is carried out for the input and output precise value variables, which are described by seven words, namely {negative large, negative medium, negative small, zero, positive small, positive medium, positive large}, so the input and output variables of fuzzy control are set as follows:

The fuzzy set of Ep is {NB, NM, NS, N0, 0, PS, PM, PB};

The fuzzy set of Cp is {NB, NM, NS, N0, 0, PS, PM, PB};

The fuzzy set of △U is {NB, NM, NS, 0, PS, PM, PB};

The domains of Ep, Cp, and △U are all {-6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6};

Step 3: Determine the membership function of the fuzzy subset. The input variable adopts a triangular membership function, which has a simple shape, less calculation workload, saves storage space, and has high control sensitivity. The output variable uses the trapezoidal membership function, which can make the fuzzy controller have good control performance on the nonlinear system.

Step four, formulate fuzzy control rules. Use If...then conditional statements to control the rules. When the error and the rate of change of the error are large or large, the selected control quantity is mainly to eliminate the error; when the error and the rate of change of the error are small, the selected control quantity should pay attention to prevent overshooting, and the stability of the system should be considered. Sex is the main consideration; when the error and the error rate of change are positive or both are negative, the corresponding sign needs to be changed to eliminate the error as the main purpose.

Step 5. Using the defuzzification method of the center of gravity method, the output fuzzy set is converted into a definite numerical output, and a fuzzy lookup table is made, and the corresponding output control quantity is obtained through the lookup table for a given input.

Step six, determine the parameters of the fuzzy controller. Determine the change interval of the input spindle power, the change interval of the spindle power change rate, and the change interval of the output feedrate override. At the same time, it is necessary to choose a reasonable scale factor and quantization factor. When the domain of error is set to [-x,+x], the fuzzy set domain of error is {-n, -n+1,...,0,...,n -1, n}, so the quantization factor K of the error is expressed as follows:

Among them, n is the maximum value of the discourse domain of the error fuzzy set, and x is the maximum value of the discourse domain of the error. Similarly, the quantization factor of the error change rate and the scaling factor of the output control quantity can also be obtained by the above method. However, the values of these control parameters need to be corrected according to the processing experiment to achieve a certain control effect.

Step seven, making fuzzy control lookup table. Make a simple fuzzy controller in the SIMULINK module of MATLAB software, and load the previously established fuzzy rules into the workspace, map the input and output variables to the corresponding input and output of the fuzzy controller in the system test, run the test and save the results, Transform the result to get the corresponding fuzzy control query table.

Step eight, replace the fuzzy controller in the online optimization process with the fuzzy control look-up table, and establish an online optimization model based on fuzzy control, which is built into the CNC machine tool through programming.

Step nine, online optimization and debugging. Collect real-time milling power data through CNC milling workpieces, use the online optimization controller to adaptively adjust the feed rate and then adjust the feed speed, and feed the adjusted power back to the constant power controller for continuous iterative adjustment until the power is constant .

Application examples.

Step 1: Collect the spindle load during the cutting process through the built-in sensor of the CNC machine tool, and multiply the spindle load by the rated power to obtain the power value of the machine tool spindle.

Step two, determine the input and output of the fuzzy controller. Given the target cutting power P _obj , compare the collected processing power with the target power value P _obj to obtain the power error E _P , and at the same time differentiate the error, and calculate the rate of change of the power error C _P in one sampling period. Therefore, The power error E _P and the power error change rate C _P are used as the input variables of the fuzzy controller. The feed override of the CNC machine tool is used as the output.

Step 3, input and output fuzzy processing. Use {negative large, negative medium, negative small, zero, positive small, positive medium, positive large} to describe the size of the fuzzy quantity, and set the input and output variables of the fuzzy control as follows:

The fuzzy set of Ep is {NB, NM, NS, N0, 0, PS, PM, PB};

The fuzzy set of Cp is {NB, NM, NS, N0, 0, PS, PM, PB};

The fuzzy set of △U is {NB, NM, NS, 0, PS, PM, PB};

The domains of Ep, Cp, and △U are all {-6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6};

Step 4. Referring to Fig. 2, it is a membership function diagram of input and output, and its input variable adopts a triangular membership function, which has a simple shape, less calculation workload, saves storage space, and has high control sensitivity. The output variable uses the trapezoidal membership function, which can make the fuzzy controller have good control performance on the nonlinear system. Use the fuzzy logic toolbox of Matlab software to complete the design of the required fuzzy logic controller, and edit the previously defined membership functions of E _P , C _P and ΔU in the fuzzy logic editing window. The two membership functions are all triangles, the output membership function is trapezoidal, and the range of the membership function is defined as [-6, 6].

Step five, formulate fuzzy control rules. The established fuzzy control rules are shown in Table 1.

① When the error is negative and the error change is positive and small, the change of the control quantity is taken as negative;

②When the error is negatively large and the error change is positively large or positively medium, the change of the control quantity is negatively small or zero level, and the control quantity should not be increased too much, otherwise it will cause overshoot and positive error;

③When the error is negative and medium, the control variable change is basically the same as when the error is negative and large, in order to eliminate the error as soon as possible; ④When the error is negative and the error change is negative, the control variable change is negative large or negative, Suppress the change of the error in the negative direction; ⑤When the error is negative and the error change is positive, the trend of the system is to eliminate the negative small error, and the change of the control variable is positive or medium.

Table 1 Fuzzy control rule table

Step 6. Defuzzification using the center of gravity method. Select "centroid" in the "Defuzification" option in the fuzzy logic toolbox, and the system will automatically use the center of gravity method to complete the defuzzification of the control quantity.

Step seven, determine the parameters of the fuzzy controller. If it is assumed that the domain of the error is set to [-x,+x], and the fuzzy set domain of the error is {-n, -n+1,...,0,...,n-1,n+1}, so the error The quantization factor K of is:

Among them, n is the maximum value of the discourse domain of the error fuzzy set, and x is the maximum value of the discourse domain of the error. Similarly, the quantization factor of the error change and the scaling factor of the output control quantity can also be obtained by the above method. According to the processing optimization, after repeated debugging, the power error quantization factor K ₁ =0.045, the power error change rate quantization factor K ₂ =0.034, and the output feed rate defuzzification quantization factor K ₃ =0.3.

Step eight, making fuzzy control lookup table. Create a simple fuzzy controller in the SIMULINK module of MATLAB software, and load the previously established fuzzy rules into the workspace. In the system test, the input and output variables are mapped to the corresponding input and output of the fuzzy controller. After running the test for 169 iterations And save the result, convert the result to get the corresponding fuzzy control query table as shown in Table 2. Create a new simple fuzzy control model in Simulink, replace the part of Fuzzy Logic Controller with Ruleviewer with complex calculations with Lookup Table, and set its parameters to the data of the fuzzy rule lookup table established in Table 2, so as to improve the calculation efficiency when performing fuzzy calculations. Efficiency, and its effect is also equivalent to the result of the above-mentioned operation with the membership function. For the operation of inputting a large amount of data, the operation time is reduced.

Table 2 Fuzzy control query table

Step 9: Establish an online optimization simulation model based on fuzzy control, and build it into the numerical control machine tool through programming. It mainly includes three parts: power simulation module, fuzzy control regulation module and power feedback output module. In order to simulate the power change of the machine tool under the condition of variable depth of cut, the input depth of cut is set as a piecewise function, and the output power is mixed with a white noise signal (White Noises), and the simulation obtains a fluctuating power signal. The fuzzy control regulation module first makes a difference between the power signal simulated by the power generator and the given target power value of 350W, and fuzzifies the power error and power error change rate through the fuzzy quantization factor, and the power feedback output module feeds the output of the fuzzy controller The magnification is fed back to the power generator to generate an optimized power signal. After the process is completed, it is fed back to the fuzzy control module until the power is constant, and the regulated feed speed value is also output. This module belongs to the process of feedback adjustment in the process of processing. Through continuous feedback and online adjustment, the constant power constraint is finally realized, so that the milling process can be carried out smoothly, the processing time can be shortened, and the processing efficiency can be improved.

It can be seen from Figure 4 that the fuzzy control-based feed speed optimization simulation model established by the above steps is used to carry out simulation experiments, and an oscilloscope is set up in each link to facilitate observation of the results of each link. It can be seen that when the given target power value is At 350W, the power at the small power difference remains basically unchanged after optimization, and the power at the large power difference approaches the target power value after optimization, and remains relatively stable, and its power fluctuation error remains at Within ±10%, this shows that the processing power after the optimized feed rate achieves a constant power constraint, making the processing process more stable, and the power after optimization is generally greater than the power before optimization, indicating that the feed speed after optimization is greater than the original feed speed , The increase of the feed speed shortens the processing time and improves the processing efficiency.

Claims (1)

1. A fuzzy control-based feed speed online optimization method is characterized by comprising the following steps:

Step one, determining input and output of a fuzzy controller; setting a target cutting power P_objCollecting the main shaft power of the machine tool in the machining process, and comparing the actual machining power with a target power value P_objcomparing to obtain power error E_PWhile differentiating the error to calculate the power error change rate C in a sampling period_PError in power E_PAnd rate of change of power error C_PAs an input variable of the fuzzy controller, taking the feed multiplying power of the numerical control machine as an output quantity;

Step two, performing input and output fuzzification processing; the input and output accurate value variable is correspondingly fuzzified and described by seven words, namely { negative large, negative medium, negative small, zero, positive small, positive medium, positive large }, and the input and output variables of the fuzzy control are set as follows:

The fuzzy set of Ep is { NB, NM, NS, N0, 0, PS, PM, PB };

The fuzzy set of Cp is { NB, NM, NS, N0, 0, PS, PM, PB };

the fuzzy set of Δ U is { NB, NM, NS, 0, PS, PM, PB };

the domains of Ep, Cp and delta U are { -6, -5, -4, -3, -2, -1, 0, 1, 2, 3, 4, 5, 6 };

Step three, determining a membership function of the fuzzy subset; the input variable adopts a triangular membership function, and the output variable adopts a trapezoidal membership function;

step four, formulating a fuzzy control rule; adopting a conditional statement control rule of If … then;

Converting the output fuzzy set into a determined numerical value for output by adopting a defuzzification method of a gravity center method, manufacturing a fuzzy lookup table, and obtaining corresponding output control quantity for given input through the lookup table;

Step six, determining parameters of a fuzzy controller; determining an input main shaft power change interval, a main shaft power change rate change interval and a feed multiplying factor change interval as an output; and simultaneously, reasonable scale factors and quantization factors are selected, when the discourse domain of the error is set to be [ -x, + x ], the fuzzy set discourse domain of the error is { -n, -n +1, …, 0, …, n-1, n }, and the quantization factor K of the error is expressed as the following formula:

Wherein n is the maximum value of the discourse domain of the error fuzzy set, and x is the maximum value of the error discourse domain;

Seventhly, manufacturing a fuzzy control look-up table; manufacturing a simple fuzzy controller in a SIMULINK module of MATLAB software, loading a previously established fuzzy rule into a working space, mapping input and output variables to corresponding input and output of the fuzzy controller in a system test, running the test, storing a result, and performing format conversion on the result to obtain a corresponding fuzzy control query table;

Replacing a fuzzy controller in the online optimization process with a fuzzy control query table, establishing an online optimization model based on fuzzy control, and programming the online optimization model to be built in a numerical control machine;

Step nine, optimizing and debugging on line; real-time milling power data are collected through a numerical control milling workpiece, the feeding multiplying power is adaptively controlled by using an online optimization controller so as to adjust the feeding speed, and the adjusted power is fed back to a constant power controller to be continuously adjusted in an iterative manner until the power is constant.